Csl

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
preview_csl.py — 用 citeproc-py 加载 .csl 文件和测试数据，渲染 citation 和 bibliography 输出。

用法:
    python preview_csl.py <file.csl>                     # 使用默认 test_data.json
    python preview_csl.py <file.csl> --data custom.json   # 使用自定义数据
"""

import argparse
import json
import os
import sys

# Windows 下强制 UTF-8 输出，避免中文乱码
if sys.platform == "win32":
    os.environ.setdefault("PYTHONUTF8", "1")
    if hasattr(sys.stdout, "reconfigure"):
        sys.stdout.reconfigure(encoding="utf-8")
    if hasattr(sys.stderr, "reconfigure"):
        sys.stderr.reconfigure(encoding="utf-8")

# ---------------------------------------------------------------------------
# 依赖检查
# ---------------------------------------------------------------------------
try:
    from citeproc import (
        Citation,
        CitationItem,
        CitationStylesBibliography,
        CitationStylesStyle,
        formatter,
    )
    from citeproc.source.json import CiteProcJSON
except ImportError:
    print("错误: 缺少 citeproc-py 库。请先安装：")
    print("  pip install citeproc-py")
    sys.exit(1)


# ---------------------------------------------------------------------------
# 辅助函数
# ---------------------------------------------------------------------------

def load_test_data(data_path: str) -> list[dict]:
    """加载 CSL-JSON 格式的测试数据。"""
    with open(data_path, "r", encoding="utf-8") as f:
        data = json.load(f)
    if not isinstance(data, list) or len(data) == 0:
        print(f"错误: {data_path} 应为非空 JSON 数组")
        sys.exit(1)
    return data


def warn_callback(citation_warning):
    """citeproc 回调，用于捕获警告（这里静默忽略）。"""
    pass


def render_citation_text(bib: CitationStylesBibliography, item_ids: list[str]) -> str:
    """注册并渲染一个 citation，返回字符串。"""
    citation = Citation([CitationItem(item_id) for item_id in item_ids])
    bib.register(citation)
    return str(bib.cite(citation, warn_callback))


def preview_csl(csl_path: str, data_path: str) -> None:
    """主流程：加载 CSL 和数据，输出 citation + bibliography。"""

    # --- 加载数据 ---
    test_data = load_test_data(data_path)
    item_ids = [entry["id"] for entry in test_data]

    # --- 加载 CSL 样式 ---
    try:
        style = CitationStylesStyle(csl_path, validate=False)
    except Exception as exc:
        print(f"错误: 无法加载 CSL 文件 '{csl_path}': {exc}")
        sys.exit(1)

    # --- 构建 bibliography ---
    source = CiteProcJSON(test_data)
    bib = CitationStylesBibliography(style, source, formatter.plain)

    # === Citation (正文引用) ===
    print("=== Citation (正文引用) ===")

    # Single: 第 1 条
    single = render_citation_text(bib, [item_ids[0]])
    print(f"Single: {single}")

    # Multiple: 第 1 条 + 第 3 条（不连续）
    if len(item_ids) >= 3:
        multiple = render_citation_text(bib, [item_ids[0], item_ids[2]])
        print(f"Multiple: {multiple}")

    # Range: 前 3 条（连续）
    if len(item_ids) >= 3:
        range_cite = render_citation_text(bib, [item_ids[0], item_ids[1], item_ids[2]])
        print(f"Range: {range_cite}")

    # --- 为剩余未注册的条目也生成 citation，确保 bibliography 包含所有条目 ---
    registered = set()
    registered.update(item_ids[:3] if len(item_ids) >= 3 else item_ids[:1])
    remaining = [iid for iid in item_ids if iid not in registered]
    if remaining:
        rest_citation = Citation([CitationItem(iid) for iid in remaining])
        bib.register(rest_citation)
        bib.cite(rest_citation, warn_callback)

    # === Bibliography (参考文献列表) ===
    print()
    print("=== Bibliography (参考文献列表) ===")
    bibliography = bib.bibliography()
    if bibliography:
        for item in bibliography:
            text = str(item).strip()
            if text:
                print(text)
    else:
        print("(bibliography 为空，请检查 CSL 文件配置)")


# ---------------------------------------------------------------------------
# CLI 入口
# ---------------------------------------------------------------------------

def main():
    parser = argparse.ArgumentParser(
        description="用 citeproc-py 预览 CSL 样式的 citation 和 bibliography 输出"
    )
    parser.add_argument("csl_file", help="CSL 样式文件路径")
    parser.add_argument(
        "--data",
        default=None,
        help="CSL-JSON 测试数据文件路径 (默认使用同目录下的 test_data.json)",
    )
    args = parser.parse_args()

    # CSL 文件
    csl_path = os.path.abspath(args.csl_file)
    if not os.path.isfile(csl_path):
        print(f"错误: CSL 文件不存在: {csl_path}")
        sys.exit(1)

    # 数据文件
    if args.data:
        data_path = os.path.abspath(args.data)
    else:
        data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data.json")
    if not os.path.isfile(data_path):
        print(f"错误: 数据文件不存在: {data_path}")
        sys.exit(1)

    print(f"CSL:  {csl_path}")
    print(f"Data: {data_path}")
    print()

    preview_csl(csl_path, data_path)


if __name__ == "__main__":
    main()

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"


## Categories for style metadata
div {
  category.citation-format =
    "author" | "author-date" | "label" | "note" | "numeric"
  
  ## Use "generic-base" for styles that are non-discipline specific, such as
  ## APA, Harvard, etc.
  category.field =
    "anthropology"
    | "astronomy"
    | "biology"
    | "botany"
    | "chemistry"
    | "communications"
    | "engineering"
    | "generic-base"
    | "geography"
    | "geology"
    | "history"
    | "humanities"
    | "law"
    | "linguistics"
    | "literature"
    | "math"
    | "medicine"
    | "philosophy"
    | "physics"
    | "political_science"
    | "psychology"
    | "science"
    | "social_science"
    | "sociology"
    | "theology"
    | "zoology"
}

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"
namespace cs = "http://purl.org/net/xbiblio/csl"


## cs:choose - Conditional Statements"
div {
  rendering-element.choose =
    
    ## Use to conditionally render rendering elements.
    element cs:choose { choose.if, choose.else-if*, choose.else? }
  choose.if = element cs:if { condition+, match, rendering-element* }
  choose.else-if =
    element cs:else-if { condition+, match, rendering-element* }
  choose.else = element cs:else { rendering-element+ }
  condition =
    
    ## If used, the element content is only rendered if it disambiguates two
    ## otherwise identical citations. This attempt at disambiguation is only
    ## made after all other disambiguation methods have failed.
    [ a:defaultValue = "true" ] attribute disambiguate { "true" }
    | 
      ## Tests whether the given variables contain numeric text.
      attribute is-numeric {
        list { variables+ }
      }
    | 
      ## Tests whether the given date variables contain approximate dates.
      attribute is-uncertain-date {
        list { variables.dates+ }
      }
    | 
      ## Tests whether the locator matches the given locator types.
      attribute locator {
        list { terms.locator+ }
      }
    | 
      ## Tests whether the cite position matches the given positions.
      attribute position {
        list {
          ("first"
           | "subsequent"
           | "ibid"
           | "ibid-with-locator"
           | "near-note")+
        }
      }
    | 
      ## Tests whether the item matches the given types.
      attribute type {
        list { item-types+ }
      }
    | 
      ## Tests whether the default ("long") forms of the given variables
      ## contain non-empty values.
      attribute variable {
        list { variables+ }
      }
  match =
    
    ## Set the testing logic.
    [ a:defaultValue = "all" ]
    attribute match {
      
      ## Element only tests "true" when all conditions test "true" for all
      ## given test values.
      "all"
      | 
        ## Element tests "true" when any condition tests "true" for any given
        ## test value.
        "any"
      | 
        ## Element only tests "true" when none of the conditions test "true"
        ## for any given test value.
        "none"
    }?
}

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"


## Terms
div {
  terms =
    terms.gender-assignable
    | terms.gender-variants
    | terms.locator
    | item-types
    | 
      ## Contributor roles
      variables.names
    | "editortranslator"
    | 
      ## Miscellaneous terms
      "accessed"
    | "ad"
    | "advance-online-publication"
    | "album"
    | "and"
    | "and others"
    | "anonymous"
    | "at"
    | "audio-recording"
    | "available at"
    | "bc"
    | "bce"
    | "by"
    | "ce"
    | "circa"
    | "cited"
    | "et-al"
    | "film"
    | "forthcoming"
    | "from"
    | "henceforth"
    | "ibid"
    | "in"
    | "in press"
    | "internet"
    | "interview"
    | "letter"
    | "loc-cit"
    | "no date"
    | "no-place"
    | "no-publisher"
    | "on"
    | "online"
    | "op-cit"
    | "original-work-published"
    | "personal-communication"
    | "podcast"
    | "podcast-episode"
    | "preprint"
    | "presented at"
    | "radio-broadcast"
    | "radio-series"
    | "radio-series-episode"
    | "reference"
    | "retrieved"
    | "review-of"
    | "scale"
    | "special-issue"
    | "special-section"
    | "television-broadcast"
    | "television-series"
    | "television-series-episode"
    | "video"
    | "working-paper"
    | 
      ## Punctuation
      "open-quote"
    | "close-quote"
    | "open-inner-quote"
    | "close-inner-quote"
    | "page-range-delimiter"
    | "colon"
    | "comma"
    | "semicolon"
    | 
      ## Seasons
      "season-01"
    | "season-02"
    | "season-03"
    | "season-04"
  
  ## Terms to which a gender may be assigned
  terms.gender-assignable =
    
    ## Months
    "month-01"
    | "month-02"
    | "month-03"
    | "month-04"
    | "month-05"
    | "month-06"
    | "month-07"
    | "month-08"
    | "month-09"
    | "month-10"
    | "month-11"
    | "month-12"
    | terms.non-locator-number-variables
    | terms.locator-number-variables
  
  ## Terms for which gender variants may be specified
  terms.gender-variants = terms.ordinals | terms.long-ordinals
  terms.ordinals =
    
    ## Ordinals
    xsd:string { pattern = "ordinal(-\d{2})?" }
  terms.long-ordinals =
    
    ## Long ordinals
    "long-ordinal-01"
    | "long-ordinal-02"
    | "long-ordinal-03"
    | "long-ordinal-04"
    | "long-ordinal-05"
    | "long-ordinal-06"
    | "long-ordinal-07"
    | "long-ordinal-08"
    | "long-ordinal-09"
    | "long-ordinal-10"
  
  ## Locators
  terms.locator =
    "act"
    | "appendix"
    | "article-locator"
    | "book"
    | "canon"
    | "chapter"
    | "column"
    | "elocation"
    | "equation"
    | "figure"
    | "folio"
    | "line"
    | "note"
    | "opus"
    | "paragraph"
    | "rule"
    | "scene"
    | "sub-verbo"
    | "table"
    | "timestamp"
    | "title-locator"
    | "verse"
    | terms.locator-number-variables
  
  ## Locator terms with matching number variables
  terms.locator-number-variables =
    "issue"
    | "page"
    | "part"
    | "section"
    | "supplement"
    | "version"
    | "volume"
  
  ## Non-locator terms accompanying number variables
  terms.non-locator-number-variables =
    "chapter-number"
    | "citation-number"
    | "collection-number"
    | "edition"
    | "first-reference-note-number"
    | "number"
    | "number-of-pages"
    | "number-of-volumes"
    | "page-first"
    | "printing"
}

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"


## Item types
div {
  item-types =
    "article"
    | "article-journal"
    | "article-magazine"
    | "article-newspaper"
    | "bill"
    | "book"
    | "broadcast"
    | "chapter"
    | "classic"
    | "collection"
    | "dataset"
    | "document"
    | "entry"
    | "entry-dictionary"
    | "entry-encyclopedia"
    | "event"
    | "figure"
    | "graphic"
    | "hearing"
    | "interview"
    | "legal_case"
    | "legislation"
    | "manuscript"
    | "map"
    | "motion_picture"
    | "musical_score"
    | "pamphlet"
    | "paper-conference"
    | "patent"
    | "performance"
    | "periodical"
    | "personal_communication"
    | "post"
    | "post-weblog"
    | "regulation"
    | "report"
    | "review"
    | "review-book"
    | "software"
    | "song"
    | "speech"
    | "standard"
    | "thesis"
    | "treaty"
    | "webpage"
}

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"


## Variables
div {
  
  ## All variables
  variables = variables.dates | variables.names | variables.standard
  
  ## Standard variables
  variables.standard =
    variables.numbers | variables.strings | variables.titles
  
  ## Date variables
  variables.dates =
    "accessed"
    | "available-date"
    | "event-date"
    | "issued"
    | "original-date"
    | "submitted"
  
  ## Name variables
  variables.names =
    "author"
    | "chair"
    | "collection-editor"
    | "compiler"
    | "composer"
    | "container-author"
    | "contributor"
    | "curator"
    | "director"
    | "editor"
    | "editor-translator"
    | "editorial-director"
    | "executive-producer"
    | "guest"
    | "host"
    | "illustrator"
    | "interviewer"
    | "narrator"
    | "organizer"
    | "original-author"
    | "performer"
    | "producer"
    | "recipient"
    | "reviewed-author"
    | "script-writer"
    | "series-creator"
    | "translator"
  
  ## Number variables
  variables.numbers =
    "chapter-number"
    | "citation-number"
    | "collection-number"
    | "edition"
    | "first-reference-note-number"
    | "issue"
    | "locator"
    | "number"
    | "number-of-pages"
    | "number-of-volumes"
    | "page"
    | "page-first"
    | "part-number"
    | "printing-number"
    | "section"
    | "supplement-number"
    | "version"
    | "volume"
  
  ## Title variables
  variables.titles =
    "collection-title"
    | "container-title"
    | "original-title"
    | "part-title"
    | "reviewed-title"
    | "title"
    | "volume-title"
    | # Short title forms. Will be removed in CSL 1.1
      "title-short"
    | "container-title-short"
  
  ## String variables
  variables.strings =
    "abstract"
    | "annote"
    | "archive"
    | "archive_collection"
    | "archive_location"
    | "archive-place"
    | "authority"
    | "call-number"
    | "citation-key"
    | "citation-label"
    | "dimensions"
    | "division"
    | "DOI"
    | # Alias for 'event-title'. Deprecated. Will be removed in CSL 1.1.
      "event"
    | "event-title"
    | "event-place"
    | "genre"
    | "ISBN"
    | "ISSN"
    | "jurisdiction"
    | "keyword"
    | "language"
    | "license"
    | "medium"
    | "note"
    | "original-publisher"
    | "original-publisher-place"
    | "PMCID"
    | "PMID"
    | "publisher"
    | "publisher-place"
    | "references"
    | "reviewed-genre"
    | "scale"
    | "source"
    | "status"
    | "URL"
    | "year-suffix"
}

namespace a = "http://relaxng.org/ns/compatibility/annotations/1.0"
namespace bibo = "http://purl.org/ontology/bibo/"
namespace cs = "http://purl.org/net/xbiblio/csl"
namespace dc = "http://purl.org/dc/elements/1.1/"
namespace sch = "http://purl.oclc.org/dsdl/schematron"
namespace xhtml = "http://www.w3.org/1999/xhtml"

# CSL schema metadata

dc:title [ "Citation Style Language" ]
dc:creator [ "Bruce D'Arcus" ]
dc:creator [ "Simon Kornblith" ]
bibo:editor [ "Frank Bennett" ]
bibo:editor [ "Rintze Zelle" ]
dc:rights [
  "Copyright 2007-2020 Citation Style Language and contributors"
]
dc:license [ "MIT license" ]
dc:description [
  "RELAX NG compact schema for the Citation Style Language (CSL)."
]

## Subparts of the CSL schema
include "csl-choose.rnc"
include "csl-terms.rnc"
include "csl-types.rnc"
include "csl-variables.rnc"
include "csl-categories.rnc"
# ==============================================================================

## cs:style and cs:locale - Root Elements
div {
  start =
    independent-style.style | dependent-style.style | locale-file.locale
  independent-style.style =
    element cs:style {
      
      ## Select whether citations appear in-text or as notes.
      attribute class { "in-text" | "note" },
      style.default-locale,
      style.options,
      version,
      independent-style.style.info,
      (style.locale*
       & style.macro*
       & style.citation
       & style.bibliography?)
    }
  dependent-style.style =
    element cs:style {
      style.default-locale, version, dependent-style.style.info
    }
  style.default-locale =
    
    ## Set a default style locale.
    attribute default-locale { xsd:language }?
  version =
    
    ## Indicate CSL version compatibility.
    [ a:defaultValue = "1.0" ] attribute version { "1.0" }
}
# ==============================================================================

## cs:info - Style and Locale File Metadata
div {
  
  ## Metadata for independent styles.
  independent-style.style.info =
    element cs:info {
      info.author*
      & info.category*
      & info.contributor*
      & info.id
      & info.issn*
      & info.eissn?
      & info.issnl?
      & independent-style.info.link*
      & info.published?
      & info.rights?
      & info.summary?
      & info.title
      & info.title-short?
      & info.updated
    }
  
  ## Metadata for dependent styles.
  dependent-style.style.info =
    element cs:info {
      info.author*
      & info.category*
      & info.contributor*
      & info.id
      & info.issn*
      & info.eissn?
      & info.issnl?
      & dependent-style.info.link+
      & info.published?
      & info.rights?
      & info.summary?
      & info.title
      & info.title-short?
      & info.updated
    }
  
  ## Metadata for locale files.
  locale-file.locale.info =
    element cs:info { info.translator* & info.rights? & info.updated? }
  info.author = element cs:author { personal-details }
  info.contributor = element cs:contributor { personal-details }
  info.translator = element cs:translator { personal-details }
  personal-details =
    element cs:name { text }
    & element cs:email { text }?
    & element cs:uri { xsd:anyURI }?
  info.category =
    
    ## Specify the citation format of the style (using the "citation-format"
    ## attribute) or the fields and disciplines for which the style is
    ## relevant (using the "field" attribute).
    element cs:category {
      attribute citation-format { category.citation-format }
      | attribute field { category.field }
    }
  info.id =
    
    ## Specify the unique and stable identifier for the style. A URI
    ## is valid, but new styles should use a UUID to ensure stability
    ## and uniqueness.
    element cs:id { xsd:anyURI }
  info.issn =
    
    ## Specify the journal's ISSN(s) for journal-specific styles. An ISSN
    ## must consist of four digits, a hyphen, three digits, and a check
    ## digit (a numeral digit or roman X), e.g. "1234-1231".
    element cs:issn { issn }
  info.eissn =
    
    ## Specify the journal's eISSN for journal-specific styles.
    element cs:eissn { issn }
  info.issnl =
    
    ## Specify the journal's ISSN-L for journal-specific styles.
    element cs:issnl { issn }
  issn = xsd:string { pattern = "\d{4}\-\d{3}(\d|x|X)" }
  independent-style.info.link =
    element cs:link {
      attribute href { xsd:anyURI },
      
      ## Specify how the URL relates to the style.
      attribute rel {
        
        ## The URI of the CSL style itself.
        "self"
        | 
          ## URI of the style from which the current style is derived.
          "template"
        | 
          ## URI of style documentation.
          "documentation"
      },
      info-text
    }
  dependent-style.info.link =
    element cs:link {
      attribute href { xsd:anyURI },
      
      ## Specify how the URL relates to the style.
      attribute rel {
        
        ## The URI of the CSL style itself.
        "self"
        | 
          ## URI of the CSL style whose content should be used for
          ## processing. Required for dependent styles.
          "independent-parent"
        | 
          ## URI of style documentation.
          "documentation"
      },
      info-text
    }
  info.published =
    
    ## Specify when the style was initially created or made available.
    element cs:published { xsd:dateTime }
  info.rights =
    element cs:rights {
      attribute license { xsd:anyURI }?,
      info-text
    }
  info.summary = element cs:summary { info-text }
  info.title = element cs:title { info-text }
  info.title-short =
    
    ## Specify an abbreviated style title (e.g., "APA")
    element cs:title-short { info-text }
  info.updated =
    
    ## Specify when the style was last updated (e.g.,
    ## "2007-10-26T21:32:52+02:00")
    element cs:updated { xsd:dateTime }
  info-text =
    attribute xml:lang { xsd:language }?,
    text
}
# ==============================================================================

## cs:locale in Independent Styles
div {
  style.locale =
    
    ## Use to (re)define localized terms, dates and options.
    element cs:locale {
      
      ## Specify the affected locale(s). If "xml:lang" is not set, the
      ## "cs:locale" element affects all locales.
      attribute xml:lang { xsd:language }?,
      (locale.style-options? & locale.date* & locale.terms?)
    }
}
# ==============================================================================

## cs:locale Contents - Localization Data
div {
  
  ## Localized global options are specified as attributes in the
  ## cs:style-options element. If future versions of CSL include localized
  ## options that are citation or bibliography specific, the elements
  ## cs:citation-options and cs:bibliography-options can be added.
  locale.style-options =
    element cs:style-options {
      
      ## Limit the "ordinal" form to the first day of the month.
      [ a:defaultValue = "false" ]
      attribute limit-day-ordinals-to-day-1 { xsd:boolean }?,
      
      ## Specify whether punctuation (a period or comma) is placed within
      ## or outside (default) the closing quotation mark.
      [ a:defaultValue = "false" ]
      attribute punctuation-in-quote { xsd:boolean }?
    }
  locale-file.locale =
    element cs:locale {
      
      ## Specify the locale of the locale file.
      attribute xml:lang { xsd:language },
      version,
      locale-file.locale.info?,
      (locale.style-options & locale.date+ & locale.terms)
    }
  locale.date =
    element cs:date {
      date.form,
      delimiter,
      font-formatting,
      text-case,
      locale.date.date-part+
    }
  date.form =
    
    ## Select the localized date format ("text" or "numeric").
    attribute form {
      
      ## Text date form (e.g., "December 15, 2005").
      "text"
      | 
        ## Numeric date form (e.g., "2005-12-15").
        "numeric"
    }
  locale.date.date-part =
    element cs:date-part {
      affixes, font-formatting, text-case, (day | month | year)
    }
  locale.terms = element cs:terms { terms.term+ }
  
  ## The "cs:term" element can either hold a basic string, or "cs:single" and
  ## "cs:multiple" child elements to give singular and plural forms of the term.
  terms.term =
    element cs:term {
      term.attributes,
      (text | (term.single, term.multiple))
    }
  term.attributes =
    (attribute name { terms },
     [ a:defaultValue = "long" ] attribute form { term.form }?)
    | (attribute name { terms.ordinals },
       attribute form { "long" }?,
       attribute gender-form { "masculine" | "feminine" }?,
       attribute match {
         "last-digit" | "last-two-digits" | "whole-number"
       }?)
    | (attribute name { terms.long-ordinals },
       attribute form { "long" }?,
       attribute gender-form { "masculine" | "feminine" })
    | (attribute name { terms.gender-assignable },
       attribute form { "long" }?,
       attribute gender { "masculine" | "feminine" })
  
  ## "verb-short" reverts to "verb" if the "verb-short" form is not available.
  ## "symbol" reverts to "short" if the "symbol" form is not available.
  ## "verb" and "short" revert to "long" if the specified form is not available.
  term.form = "long" | "verb" | "short" | "verb-short" | "symbol"
  term.single =
    
    ## Singular version of the term.
    element cs:single { text }
  term.multiple =
    
    ## Plural version of the term.
    element cs:multiple { text }
}
# ==============================================================================

## cs:macro
div {
  style.macro =
    
    ## Use to create collections of (reusable) formatting instructions.
    element cs:macro {
      attribute name { xsd:NMTOKEN },
      rendering-element+
    }
}
# ==============================================================================

## Rendering Elements
div {
  rendering-element =
    rendering-element.names
    | rendering-element.date
    | rendering-element.label
    | rendering-element.text
    | rendering-element.number
    | rendering-element.choose
    | rendering-element.group
}
# ==============================================================================

## cs:citation and cs:bibliography
div {
  style.citation =
    
    ## Use to describe the formatting of citations.
    element cs:citation { citation.options, sort?, citation.layout }
  style.bibliography =
    
    ## Use to describe the formatting of the bibliography.
    element cs:bibliography {
      bibliography.options, sort?, bibliography.layout
    }
  citation.layout =
    element cs:layout {
      affixes, delimiter, font-formatting, rendering-element+
    }
  bibliography.layout =
    element cs:layout { affixes, font-formatting, rendering-element+ }
}
# ==============================================================================

## cs:names Rendering Element
div {
  rendering-element.names =
    element cs:names {
      names.attributes,
      ((names.name?, names.et-al?) & names.label?),
      names.substitute?
    }
  names.attributes =
    attribute variable {
      list { variables.names+ }
    },
    affixes,
    
    ## Specify the delimiter for name lists of name variables rendered by
    ## the same cs:names element.
    delimiter,
    display,
    font-formatting
  names.name =
    element cs:name {
      name.attributes,
      
      ## Select the "long" (first name + last name, for Western names),
      ## "short" (last name only, for Western names), or "count" name form
      ## (returning the number of names in the name variable, which can be
      ## useful for some sorting algorithms).
      [ a:defaultValue = "long" ]
      attribute form { "long" | "short" | "count" }?,
      affixes,
      
      ## Set the delimiter for names in a name variable (e.g., ", " in
      ## "Doe, Smith")
      [ a:defaultValue = ", " ] delimiter,
      font-formatting,
      name.name-part*
    }
  name.attributes =
    
    ## Use to separate the second-to-last and last name of a name list by
    ## the "and" term or ampersand.
    attribute and {
      
      ## Use the "and" term (e.g., "Doe, Johnson and Smith").
      "text"
      | 
        ## Use the "ampersand" (e.g., "Doe, Johnson & Smith").
        "symbol"
    }?,
    
    ## Specify when the name delimiter is used between a truncated name list
    ## and the "et-al" (or "and others") term in case of et-al abbreviation
    ## (e.g., "Smith, Doe et al." or "Smith, Doe, et al.").
    [ a:defaultValue = "contextual" ]
    attribute delimiter-precedes-et-al {
      
      ## The name delimiter is only used when the truncated name list
      ## consists of two or more names.
      "contextual"
      | 
        ## The name delimiter is always used.
        "always"
      | 
        ## The name delimiter is never used.
        "never"
      | 
        ## The name delimiter is only used if the preceding name is inverted as
        ## a result of the "name-as-sort-order" attribute.
        "after-inverted-name"
    }?,
    
    ## Specify when the name delimiter is used between the second-to-last
    ## and last name of a non-truncated name list. Only has an effect when
    ## the "and" term or ampersand is used (e.g., "Doe and Smith" or "Doe,
    ## and Smith").
    [ a:defaultValue = "contextual" ]
    attribute delimiter-precedes-last {
      
      ## The name delimiter is only used when the name list consists of
      ## three or more names.
      "contextual"
      | 
        ## The name delimiter is always used.
        "always"
      | 
        ## The name delimiter is never used.
        "never"
      | 
        ## The name delimiter is only used if the preceding name is inverted as
        ## a result of the "name-as-sort-order" attribute.
        "after-inverted-name"
    }?,
    
    ## Set the minimum number of names needed in a name variable to activate
    ## et-al abbreviation.
    attribute et-al-min { xsd:integer }?,
    
    ## Set the number of names to render when et-al abbreviation is active.
    attribute et-al-use-first { xsd:integer }?,
    
    ## As "et-al-min", but only affecting subsequent citations to an item.
    attribute et-al-subsequent-min { xsd:integer }?,
    
    ## As "et-al-use-first", but only affecting subsequent citations to an
    ## item.
    attribute et-al-subsequent-use-first { xsd:integer }?,
    
    ## If set to "true", the "et-al" (or "and others") term is replaced by
    ## an ellipsis followed by the last name of the name variable.
    [ a:defaultValue = "false" ]
    attribute et-al-use-last { xsd:boolean }?,
    
    ## If set to "false", names are not initialized and "initialize-with"
    ## only affects initials already present in the input data.
    [ a:defaultValue = "true" ] attribute initialize { xsd:boolean }?,
    
    ## Activate initializing of given names. The attribute value is appended
    ## to each initial (e.g., with ". ", "Orson Welles" becomes "O. Welles").
    attribute initialize-with { text }?,
    
    ## Specify whether (and which) names should be rendered in their sort
    ## order (e.g., "Doe, John" instead of "John Doe").
    attribute name-as-sort-order {
      
      ## Render the first name of each name variable in sort order.
      "first"
      | 
        ## Render all names in sort order.
        "all"
    }?,
    
    ## Sets the delimiter for name-parts that have switched positions as a
    ## result of "name-as-sort-order" (e.g., ", " in "Doe, John").
    [ a:defaultValue = ", " ] attribute sort-separator { text }?
  name.name-part =
    
    ## Use to format individual name parts (e.g., "Jane DOE").
    element cs:name-part {
      attribute name { "family" | "given" },
      affixes,
      font-formatting,
      text-case
    }
  names.et-al =
    
    ## Specify the term used for et-al abbreviation and its formatting.
    element cs:et-al {
      
      ## Select the term to use for et-al abbreviation.
      [ a:defaultValue = "et-al" ]
      attribute term { "et-al" | "and others" }?,
      font-formatting
    }
  
  ## Inherits variable from the parent cs:names element.
  names.label =
    element cs:label {
      [ a:defaultValue = "long" ] attribute form { term.form }?,
      label.attributes-shared
    }
  names.substitute =
    
    ## Specify substitution options when the name variables selected on the
    ## parent cs:names element are empty.
    element cs:substitute { (substitute.names | rendering-element)+ }
  
  ## Short version of cs:names, without children, allowed in cs:substitute.
  substitute.names = element cs:names { names.attributes }
}
# ==============================================================================

## cs:date Rendering Element
div {
  rendering-element.date =
    element cs:date {
      attribute variable { variables.dates },
      ((
        ## Limit the date parts rendered.
        [ a:defaultValue = "year-month-day" ]
        attribute date-parts {
          
          ## Year, month and day
          "year-month-day"
          | 
            ## Year and month
            "year-month"
          | 
            ## Year only
            "year"
        }?,
        date.form,
        rendering-element.date.date-part.localized*)
       | (rendering-element.date.date-part.non-localized+, delimiter)),
      affixes,
      display,
      font-formatting,
      text-case
    }
  rendering-element.date.date-part.localized =
    
    ## Specify overriding formatting for localized dates (affixes
    ## cannot be overridden, as these are considered locale-specific).
    ## Example uses are forcing the use of leading-zeros, or of the
    ## "short" month form. Has no effect on which, and in what order,
    ## date parts are rendered.
    element cs:date-part {
      font-formatting, text-case, (day | month | year)
    }
  rendering-element.date.date-part.non-localized =
    
    ## Specify, in the desired order, the date parts that should be
    ## rendered and their formatting.
    element cs:date-part {
      affixes, font-formatting, text-case, (day | month | year)
    }
  day =
    attribute name { "day" },
    
    ## Day forms: "numeric" ("5"), "numeric-leading-zeros" ("05"), "ordinal"
    ## ("5th").
    [ a:defaultValue = "numeric" ]
    attribute form { "numeric" | "numeric-leading-zeros" | "ordinal" }?,
    range-delimiter
  month =
    attribute name { "month" },
    
    ## Months forms: "long" (e.g., "January"), "short" ("Jan."), "numeric"
    ## ("1"), and "numeric-leading-zeros" ("01").
    [ a:defaultValue = "long" ]
    attribute form {
      "long" | "short" | "numeric" | "numeric-leading-zeros"
    }?,
    range-delimiter,
    strip-periods
  year =
    attribute name { "year" },
    
    ## Year forms: "long" ("2005"), "short" ("05").
    [ a:defaultValue = "long" ] attribute form { "short" | "long" }?,
    range-delimiter
  range-delimiter =
    
    ## Specify a delimiter for date ranges (by default the en-dash). A custom
    ## delimiter is retrieved from the largest date part ("day", "month" or
    ## "year") that differs between the two dates.
    [ a:defaultValue = "–" ] attribute range-delimiter { text }?
}
# ==============================================================================

## cs:text Rendering Element
div {
  rendering-element.text =
    
    ## Use to call macros, render variables, terms, or verbatim text.
    element cs:text {
      text.attributes,
      affixes,
      display,
      font-formatting,
      quotes,
      strip-periods,
      text-case
    }
  text.attributes =
    
    ## Select a macro.
    attribute macro { xsd:NMTOKEN }
    | (
       ## Select a term.
       attribute term { terms },
       [ a:defaultValue = "long" ] attribute form { term.form }?,
       
       ## Specify term plurality: singular ("false") or plural ("true").
       [ a:defaultValue = "false" ] attribute plural { xsd:boolean }?)
    | 
      ## Specify verbatim text.
      attribute value { text }
    | (
       ## Select a variable.
       attribute variable { variables.standard },
       [ a:defaultValue = "long" ] attribute form { "short" | "long" }?)
}
# ==============================================================================

## cs:number Rendering Element
div {
  rendering-element.number =
    
    ## Use to render a number variable.
    element cs:number {
      number.attributes, affixes, display, font-formatting, text-case
    }
  number.attributes =
    attribute variable { variables.numbers },
    
    ## Number forms: "numeric" ("4"), "ordinal" ("4th"), "long-ordinal"
    ## ("fourth"), "roman" ("iv").
    [ a:defaultValue = "numeric" ]
    attribute form { "numeric" | "ordinal" | "long-ordinal" | "roman" }?
}
# ==============================================================================

## cs:label Rendering Element
div {
  rendering-element.label =
    
    ## Use to render a term whose pluralization depends on the content of a
    ## variable. E.g., if "page" variable holds a range, the plural label
    ## "pp." is selected instead of the singular "p.".
    element cs:label { label.attributes, label.attributes-shared }
  label.attributes =
    attribute variable { variables.numbers | "locator" | "page" },
    [ a:defaultValue = "long" ]
    attribute form { "long" | "short" | "symbol" }?
  label.attributes-shared =
    
    ## Specify when the plural version of a term is selected.
    [ a:defaultValue = "contextual" ]
    attribute plural { "always" | "never" | "contextual" }?,
    affixes,
    font-formatting,
    strip-periods,
    text-case
}
# ==============================================================================

## cs:group Rendering Element
div {
  rendering-element.group =
    
    ## Use to group rendering elements. Groups are useful for setting a
    ## delimiter for the group children, for organizing the layout of
    ## bibliographic entries (using the "display" attribute), and for
    ## suppressing the rendering of terms and verbatim text when variables
    ## are empty.
    element cs:group {
      group.attributes,
      affixes,
      delimiter,
      display,
      font-formatting,
      rendering-element+
    }
  group.attributes = notAllowed?
}
# ==============================================================================

## Style Options
div {
  style.options =
    style.demote-non-dropping-particle,
    style.initialize-with-hyphen,
    style.page-range-format,
    names-inheritable-options,
    name-inheritable-options
  citation.options =
    citation.cite-group-delimiter,
    citation.collapse-options,
    citation.disambiguate-options,
    citation.near-note-distance,
    names-inheritable-options,
    name-inheritable-options
  bibliography.options =
    bibliography.hanging-indent,
    bibliography.line-formatting-options,
    bibliography.second-field-align,
    bibliography.subsequent-author-substitute-options,
    names-inheritable-options,
    name-inheritable-options
  style.demote-non-dropping-particle =
    
    ## Specify whether the non-dropping particle is demoted in inverted
    ## names (e.g., "Koning, W. de").
    [ a:defaultValue = "display-and-sort" ]
    attribute demote-non-dropping-particle {
      "never" | "sort-only" | "display-and-sort"
    }?
  style.initialize-with-hyphen =
    
    ## Specify whether compound given names (e.g., "Jean-Luc") are
    ## initialized with ("J-L") or without a hyphen ("JL").
    [ a:defaultValue = "true" ]
    attribute initialize-with-hyphen { xsd:boolean }?
  style.page-range-format =
    
    ## Reformat page ranges in the "page" variable.
    attribute page-range-format {
      "expanded"
      | "minimal"
      | "minimal-two"
      | "chicago"
      | "chicago-15"
      | "chicago-16"
    }?
  citation.cite-group-delimiter =
    
    ## Activate cite grouping and specify the delimiter for cites within a
    ## cite group.
    [ a:defaultValue = ", " ] attribute cite-group-delimiter { text }?
  citation.collapse-options =
    
    ## Activate cite grouping and specify the method of citation collapsing.
    attribute collapse {
      
      ## Collapse ranges of numeric cites, e.g. from "[1,2,3]" to "[1-3]".
      "citation-number"
      | 
        ## Collapse cites by suppressing repeated names, e.g. from "(Doe
        ## 2000, Doe 2001)" to "(Doe 2000, 2001)".
        "year"
      | 
        ## Collapse cites as with "year", but also suppresses repeated
        ## years, e.g. from "(Doe 2000a, Doe 2000b)" to "(Doe 2000a, b)".
        "year-suffix"
      | 
        ## Collapses cites as with "year-suffix", but also collapses
        ## ranges of year-suffixes, e.g. from "(Doe 2000a, Doe 2000b,
        ## Doe 2000c)" to "(Doe 2000a-c)".
        "year-suffix-ranged"
    }?,
    
    ## Specify the delimiter between year-suffixes. Defaults to the cite
    ## delimiter.
    attribute year-suffix-delimiter { text }?,
    
    ## Specify the delimiter following a group of collapsed cites. Defaults
    ## to the cite delimiter.
    attribute after-collapse-delimiter { text }?
  citation.disambiguate-options =
    
    ## Set to "true" to activate disambiguation by showing names that were
    ## originally hidden as a result of et-al abbreviation.
    [ a:defaultValue = "false" ]
    attribute disambiguate-add-names { xsd:boolean }?,
    
    ## Set to "true" to activate disambiguation by expanding names, showing
    ## initials or full given names.
    [ a:defaultValue = "false" ]
    attribute disambiguate-add-givenname { xsd:boolean }?,
    
    ## Set to "true" to activate disambiguation by adding year-suffixes
    ## (e.g., "(Doe 2007a, Doe 2007b)") for items from the same author(s)
    ## and year.
    [ a:defaultValue = "false" ]
    attribute disambiguate-add-year-suffix { xsd:boolean }?,
    
    ## Specify how name are expanded for disambiguation.
    [ a:defaultValue = "by-cite" ]
    attribute givenname-disambiguation-rule {
      
      ## Each ambiguous names is progressively transformed until
      ## disambiguated (when disambiguation is not possible, the name
      ## remains in its original form).
      "all-names"
      | 
        ## As "all-names", but name expansion is limited to showing
        ## initials.
        "all-names-with-initials"
      | 
        ## As "all-names", but disambiguation is limited to the first name
        ## of each cite.
        "primary-name"
      | 
        ## As "all-names-with-initials", but disambiguation is limited to
        ## the first name of each cite.
        "primary-name-with-initials"
      | 
        ## As "all-names", but only ambiguous names in ambiguous cites are
        ## expanded.
        "by-cite"
    }?
  citation.near-note-distance =
    
    ## Set the number of preceding notes (footnotes or endnotes) within
    ## which the current item needs to have been previously cited in order
    ## for the "near-note" position to be "true".
    [ a:defaultValue = "5" ]
    attribute near-note-distance { xsd:integer }?
  bibliography.hanging-indent =
    
    ## Set to "true" to render bibliographic entries with hanging indents.
    [ a:defaultValue = "false" ]
    attribute hanging-indent { xsd:boolean }?
  bibliography.line-formatting-options =
    
    ## Set the spacing between bibliographic entries.
    [ a:defaultValue = "1" ]
    attribute entry-spacing { xsd:nonNegativeInteger }?,
    
    ## Set the spacing between bibliographic lines.
    [ a:defaultValue = "1" ]
    attribute line-spacing {
      xsd:integer { minExclusive = "0" }
    }?
  bibliography.second-field-align =
    
    ## Use to align any subsequent lines of bibliographic entries with the
    ## beginning of the second field.
    attribute second-field-align {
      
      ## Align the first field with the margin.
      "flush"
      | 
        ## Put the first field in the margin and align all subsequent
        ## lines of text with the margin.
        "margin"
    }?
  bibliography.subsequent-author-substitute-options =
    
    ## Substitute names that repeat in subsequent bibliographic entries by
    ## the attribute value.
    attribute subsequent-author-substitute { text }?,
    
    ## Specify the method of substitution of names repeated in subsequent
    ## bibliographic entries.
    [ a:defaultValue = "complete-all" ]
    attribute subsequent-author-substitute-rule {
      
      ## Requires a match of all rendered names in the name variable, and
      ## substitutes once for all names.
      "complete-all"
      | 
        ## Requires a match of all rendered names in the name variable,
        ## and substitutes for each name.
        "complete-each"
      | 
        ## Substitutes for each name, until the first mismatch.
        "partial-each"
      | 
        ## Substitutes the first name if it matches.
        "partial-first"
    }?
  
  ## Options affecting cs:names, for cs:style, cs:citation and cs:bibliography.
  names-inheritable-options =
    
    ## Inheritable name option, companion for "delimiter" on cs:names.
    attribute names-delimiter { text }?
  
  ## Options affecting cs:name, for cs:style, cs:citation and cs:bibliography.
  name-inheritable-options =
    name.attributes,
    
    ## Inheritable name option, companion for "delimiter" on cs:name.
    attribute name-delimiter { text }?,
    
    ## Inheritable name option, companion for "form" on cs:name.
    [ a:defaultValue = "long" ]
    attribute name-form { "long" | "short" | "count" }?
}
# ==============================================================================

## cs:sort - Sorting
div {
  sort =
    
    ## Specify how cites and bibliographic entries should be sorted. By
    ## default, items appear in the order in which they were cited.
    element cs:sort { sort.key+ }
  sort.key =
    element cs:key {
      (attribute variable { variables }
       | attribute macro { xsd:NMTOKEN }),
      
      ## The minimum number of names needed in a name variable to activate
      ## name list truncation. Overrides the values set on any
      ## "et-al-(subsequent-)min" attributes.
      attribute names-min { xsd:integer }?,
      
      ## The number of names to render when name list truncation is
      ## activated. Overrides the values set on the
      ## "et-al-(subsequent-)use-first" attributes.
      attribute names-use-first { xsd:integer }?,
      
      ## Use to override the value of the "et-at-use-last" attribute.
      attribute names-use-last { xsd:boolean }?,
      
      ## Select between an ascending and descending sort.
      [ a:defaultValue = "ascending" ]
      attribute sort { "ascending" | "descending" }?
    }
}
# ==============================================================================

## Formatting attributes.
div {
  affixes =
    [ a:defaultValue = "" ] attribute prefix { text }?,
    [ a:defaultValue = "" ] attribute suffix { text }?
  delimiter = attribute delimiter { text }?
  display =
    
    ## By default, bibliographic entries consist of continuous runs of text.
    ## With the "display" attribute, portions of each entry can be
    ## individually positioned.
    attribute display {
      
      ## Places the content in a block stretching from margin to margin.
      "block"
      | 
        ## Places the content in a block starting at the left margin.
        "left-margin"
      | 
        ## Places the content in a block to the right of a preceding
        ## "left-margin" block.
        "right-inline"
      | 
        ## Places the content in a block indented to the right by a standard
        ## amount.
        "indent"
    }?
  
  ## The font-formatting attributes are based on those of CSS and XSL-FO.
  font-formatting =
    [ a:defaultValue = "normal" ]
    attribute font-style { "italic" | "normal" | "oblique" }?,
    [ a:defaultValue = "normal" ]
    attribute font-variant { "normal" | "small-caps" }?,
    [ a:defaultValue = "normal" ]
    attribute font-weight { "normal" | "bold" | "light" }?,
    [ a:defaultValue = "none" ]
    attribute text-decoration { "none" | "underline" }?,
    [ a:defaultValue = "baseline" ]
    attribute vertical-align { "baseline" | "sup" | "sub" }?
  quotes =
    
    ## When set to "true", quotes are placed around the rendered text.
    [ a:defaultValue = "false" ] attribute quotes { xsd:boolean }?
  strip-periods =
    
    ## When set to "true", periods are removed from the rendered text.
    [ a:defaultValue = "false" ]
    attribute strip-periods { xsd:boolean }?
  text-case =
    attribute text-case {
      
      ## Renders text in lowercase.
      "lowercase"
      | 
        ## Renders text in uppercase.
        "uppercase"
      | 
        ## Capitalizes the first character (other characters remain in
        ## their original case).
        "capitalize-first"
      | 
        ## Capitalizes the first character of every word (other characters
        ## remain in their original case).
        "capitalize-all"
      | 
        ## Renders text in title case.
        "title"
      | 
        ## Renders text in sentence case.
        ## Deprecated. Will be removed in CSL 1.1
        "sentence"
    }?
}

[
  {
    "id": "zhang2024novel",
    "type": "article-journal",
    "title": "A novel approach to deep reinforcement learning for autonomous systems",
    "author": [
      {"family": "Zhang", "given": "Wei"},
      {"family": "Li", "given": "Ming"},
      {"family": "Wang", "given": "Fang"},
      {"family": "Chen", "given": "Xiaohui"}
    ],
    "container-title": "IEEE Transactions on Neural Networks and Learning Systems",
    "volume": "15",
    "issue": "3",
    "page": "245-260",
    "issued": {"date-parts": [[2024]]},
    "DOI": "10.1109/TNNLS.2024.0012345"
  },
  {
    "id": "smith2023intro",
    "type": "book",
    "title": "Introduction to Machine Learning: Theory and Applications",
    "author": [
      {"family": "Smith", "given": "John"}
    ],
    "edition": "3",
    "publisher": "Academic Press",
    "publisher-place": "New York",
    "issued": {"date-parts": [[2023]]},
    "ISBN": "978-0-12-345678-9"
  },
  {
    "id": "johnson2024transfer",
    "type": "paper-conference",
    "title": "Transfer learning with large-scale pre-trained models",
    "author": [
      {"family": "Johnson", "given": "Alice"}
    ],
    "container-title": "Proceedings of the 41st International Conference on Machine Learning (ICML 2024)",
    "publisher-place": "Vienna, Austria",
    "page": "112-120",
    "issued": {"date-parts": [[2024]]}
  },
  {
    "id": "brown2023ethical",
    "type": "chapter",
    "title": "Ethical considerations in artificial intelligence deployment",
    "author": [
      {"family": "Brown", "given": "David"}
    ],
    "container-title": "Handbook of Artificial Intelligence",
    "publisher": "Springer",
    "publisher-place": "Berlin",
    "page": "45-78",
    "issued": {"date-parts": [[2023]]}
  },
  {
    "id": "who2024global",
    "type": "webpage",
    "title": "Global health report 2024",
    "author": [
      {"literal": "World Health Organization"}
    ],
    "URL": "https://www.who.int/report2024",
    "issued": {"date-parts": [[2024]]},
    "accessed": {"date-parts": [[2024, 6, 15]]}
  },
  {
    "id": "lee2023ml",
    "type": "thesis",
    "title": "Machine Learning in Healthcare: Predictive Models for Clinical Decision Support",
    "author": [
      {"family": "Lee", "given": "Sarah"}
    ],
    "publisher": "Massachusetts Institute of Technology",
    "issued": {"date-parts": [[2023]]},
    "genre": "Ph.D. Dissertation"
  },
  {
    "id": "garcia2024ai",
    "type": "article-newspaper",
    "title": "AI Breakthrough Announced: New Model Surpasses Human Performance",
    "author": [
      {"family": "Garcia", "given": "Maria"}
    ],
    "container-title": "The New York Times",
    "issued": {"date-parts": [[2024, 3, 15]]}
  },
  {
    "id": "noaa2024climate",
    "type": "report",
    "title": "Annual Climate Assessment 2024",
    "author": [
      {"literal": "National Oceanic and Atmospheric Administration"}
    ],
    "publisher": "NOAA",
    "publisher-place": "Washington, DC",
    "issued": {"date-parts": [[2024]]}
  },
  {
    "id": "zhang2024zhongyao",
    "type": "article-journal",
    "title": "基于深度学习的中文文本情感分析研究",
    "author": [
      {"family": "张", "given": "伟"},
      {"family": "李", "given": "娜"},
      {"family": "王", "given": "强"}
    ],
    "container-title": "计算机学报",
    "volume": "47",
    "issue": "5",
    "page": "1023-1035",
    "issued": {"date-parts": [[2024]]},
    "language": "zh-CN"
  },
  {
    "id": "liu2023rengong",
    "type": "book",
    "title": "人工智能导论",
    "author": [
      {"family": "刘", "given": "明"},
      {"family": "陈", "given": "静"}
    ],
    "publisher": "清华大学出版社",
    "publisher-place": "北京",
    "issued": {"date-parts": [[2023]]},
    "language": "zh-CN"
  }
]

#!/usr/bin/env python3
"""CSL file validator with three validation stages.

Usage:
    python validate_csl.py <file.csl>
    python validate_csl.py --verbose <file.csl>

Stages:
    1. XML Syntax check (well-formed, UTF-8)
    2. CSL RelaxNG Schema validation (downloads & caches schema)
    3. Logic rules check (R1-R6)
"""

import argparse
import json
import os
import re
import sys
import urllib.request
import urllib.error
from pathlib import Path

from lxml import etree

# CSL namespace
CSL_NS = "http://purl.org/net/xbiblio/csl"
NS = {"csl": CSL_NS}

# Schema download URLs (raw GitHub content)
# Use v1.0.2 tag for CSL 1.0 files (the vast majority of existing styles)
# Note: The official schema doesn't fully cover all CSL 1.0.2 features
# (e.g., multi-layout with locale attributes). Schema errors for such
# features are expected and should be interpreted with this in mind.
SCHEMA_BRANCHES = {
    "1.0": "v1.0.2",
    "1.1": "master",  # master tracks the upcoming 1.1 spec
}
SCHEMA_BASE_URL_TEMPLATE = "https://raw.githubusercontent.com/citation-style-language/schema/{branch}/schemas/styles"
SCHEMA_FILES = [
    "csl.rnc",
    "csl-categories.rnc",
    "csl-choose.rnc",
    "csl-terms.rnc",
    "csl-types.rnc",
    "csl-variables.rnc",
]

SCRIPT_DIR = Path(__file__).resolve().parent
SCHEMA_DIR = SCRIPT_DIR / "schema"


def log_verbose(msg, verbose=False):
    if verbose:
        print(f"  [verbose] {msg}", file=sys.stderr)


# ---------------------------------------------------------------------------
# Stage 1: XML Syntax Check
# ---------------------------------------------------------------------------

def stage1_xml_syntax(filepath, verbose=False):
    """Check well-formed XML and UTF-8 encoding."""
    errors = []

    # Check encoding declaration
    try:
        with open(filepath, "rb") as f:
            raw = f.read(200)
        # Try to detect encoding from XML declaration
        header = raw.decode("ascii", errors="replace")
        if "encoding=" in header.lower():
            match = re.search(r'encoding=["\']([^"\']+)["\']', header, re.IGNORECASE)
            if match:
                declared = match.group(1).lower().replace("-", "")
                if declared not in ("utf8",):
                    errors.append({
                        "rule": "XML",
                        "message": f"Encoding declared as '{match.group(1)}', expected 'utf-8'",
                        "severity": "error",
                    })
        # Try parsing as UTF-8
        with open(filepath, "r", encoding="utf-8") as f:
            f.read()
    except UnicodeDecodeError as e:
        errors.append({
            "rule": "XML",
            "message": f"File is not valid UTF-8: {e}",
            "severity": "error",
        })

    # Parse XML
    try:
        parser = etree.XMLParser(recover=False)
        tree = etree.parse(filepath, parser)
        log_verbose("XML parsed successfully", verbose)
    except etree.XMLSyntaxError as e:
        errors.append({
            "rule": "XML",
            "message": f"XML syntax error: {e}",
            "severity": "error",
        })
        return {"stage": 1, "name": "XML Syntax", "passed": len(errors) == 0, "errors": errors}, None

    passed = len(errors) == 0
    return {"stage": 1, "name": "XML Syntax", "passed": passed, "errors": errors}, tree


# ---------------------------------------------------------------------------
# Stage 2: CSL Schema Validation (RelaxNG)
# ---------------------------------------------------------------------------

def _get_schema_dir(version):
    """Return version-specific schema cache directory."""
    branch = SCHEMA_BRANCHES.get(version, SCHEMA_BRANCHES["1.0"])
    return SCHEMA_DIR / branch


def download_schema_files(version="1.0", verbose=False):
    """Download CSL .rnc schema files and cache them locally."""
    branch = SCHEMA_BRANCHES.get(version, SCHEMA_BRANCHES["1.0"])
    base_url = SCHEMA_BASE_URL_TEMPLATE.format(branch=branch)
    schema_dir = _get_schema_dir(version)
    schema_dir.mkdir(parents=True, exist_ok=True)
    log_verbose(f"Schema branch: {branch} (CSL version {version})", verbose)
    for fname in SCHEMA_FILES:
        target = schema_dir / fname
        if target.exists():
            log_verbose(f"Schema file cached: {fname}", verbose)
            continue
        url = f"{base_url}/{fname}"
        log_verbose(f"Downloading {url}", verbose)
        try:
            req = urllib.request.Request(url, headers={"User-Agent": "CSL-Validator/1.0"})
            with urllib.request.urlopen(req, timeout=30) as resp:
                data = resp.read()
            target.write_bytes(data)
            log_verbose(f"Saved {fname} ({len(data)} bytes)", verbose)
        except (urllib.error.URLError, urllib.error.HTTPError, OSError) as e:
            raise RuntimeError(f"Failed to download {fname}: {e}")


def convert_rnc_to_rng(version="1.0", verbose=False):
    """Convert csl.rnc to csl.rng using rnc2rng library. Returns path or None."""
    schema_dir = _get_schema_dir(version)
    rng_path = schema_dir / "csl.rng"
    rnc_path = schema_dir / "csl.rnc"

    if rng_path.exists():
        log_verbose("Using cached csl.rng", verbose)
        return rng_path

    try:
        import rnc2rng
    except ImportError:
        return None

    log_verbose("Converting csl.rnc -> csl.rng via rnc2rng", verbose)
    try:
        # rnc2rng.load() resolves include directives relative to cwd,
        # so we need to change to the schema directory temporarily.
        old_cwd = os.getcwd()
        os.chdir(str(schema_dir))
        try:
            tree = rnc2rng.load(str(rnc_path))
            rng_xml = rnc2rng.dumps(tree)
        finally:
            os.chdir(old_cwd)
        rng_path.write_text(rng_xml, encoding="utf-8")
        log_verbose("Conversion successful", verbose)
        return rng_path
    except Exception as e:
        log_verbose(f"rnc2rng conversion failed: {e}", verbose)
        return None


def stage2_schema_validation(tree, verbose=False):
    """Validate against CSL RelaxNG schema."""
    if tree is None:
        return {
            "stage": 2, "name": "CSL Schema", "passed": False,
            "errors": [{"rule": "Schema", "message": "Skipped: XML parse failed in Stage 1", "severity": "error"}],
            "skipped": True,
        }

    # Detect CSL version from the root element
    root = tree.getroot()
    csl_version = root.get("version", "1.0")
    # Normalize: "1.0" stays "1.0", "1.1" stays "1.1", anything else defaults to "1.0"
    if csl_version not in SCHEMA_BRANCHES:
        log_verbose(f"Unknown CSL version '{csl_version}', defaulting to 1.0 schema", verbose)
        csl_version = "1.0"

    # Download schema files
    try:
        download_schema_files(version=csl_version, verbose=verbose)
    except RuntimeError as e:
        msg = f"Schema download failed: {e}"
        log_verbose(msg, verbose)
        print(f"  WARNING: {msg}", file=sys.stderr)
        return {
            "stage": 2, "name": "CSL Schema", "passed": False,
            "errors": [{"rule": "Schema", "message": msg, "severity": "warning"}],
            "skipped": True,
        }

    # Convert .rnc to .rng
    rng_path = convert_rnc_to_rng(version=csl_version, verbose=verbose)
    if rng_path is None:
        msg = "rnc2rng library not available; cannot convert .rnc to .rng. Install with: pip install rnc2rng"
        log_verbose(msg, verbose)
        print(f"  WARNING: {msg}", file=sys.stderr)
        return {
            "stage": 2, "name": "CSL Schema", "passed": False,
            "errors": [{"rule": "Schema", "message": msg, "severity": "warning"}],
            "skipped": True,
        }

    # Validate
    errors = []
    try:
        rng_doc = etree.parse(str(rng_path))
        rng_schema = etree.RelaxNG(rng_doc)
        valid = rng_schema.validate(tree)
        if not valid:
            for err in rng_schema.error_log:
                errors.append({
                    "rule": "Schema",
                    "message": str(err),
                    "severity": "error",
                })
        log_verbose(f"Schema validation: {'PASS' if valid else 'FAIL'}", verbose)
    except etree.RelaxNGParseError as e:
        msg = f"Failed to parse RelaxNG schema: {e}"
        log_verbose(msg, verbose)
        print(f"  WARNING: {msg}", file=sys.stderr)
        return {
            "stage": 2, "name": "CSL Schema", "passed": False,
            "errors": [{"rule": "Schema", "message": msg, "severity": "warning"}],
            "skipped": True,
        }

    passed = len(errors) == 0
    return {"stage": 2, "name": "CSL Schema", "passed": passed, "errors": errors, "skipped": False}


# ---------------------------------------------------------------------------
# Stage 3: Logic Rules
# ---------------------------------------------------------------------------

def stage3_logic_rules(tree, verbose=False):
    """Custom logic rules R1-R6."""
    if tree is None:
        return {
            "stage": 3, "name": "Logic Rules", "passed": False,
            "errors": [{"rule": "R0", "message": "Skipped: XML parse failed in Stage 1", "severity": "error"}],
        }

    root = tree.getroot()
    errors = []

    # ---- R1: Structure completeness ----
    _check_r1(root, errors, verbose)

    # ---- R2: Macro reference integrity ----
    _check_r2(root, errors, verbose)

    # ---- R3: Class consistency ----
    _check_r3(root, errors, verbose)

    # ---- R4: et-al parameter validity ----
    _check_r4(root, errors, verbose)

    # ---- R5: Bilingual layout ordering ----
    _check_r5(root, errors, verbose)

    # ---- R6: No residual placeholders or empty macros ----
    _check_r6(root, errors, verbose)

    has_error = any(e["severity"] == "error" for e in errors)
    return {"stage": 3, "name": "Logic Rules", "passed": not has_error, "errors": errors}


def _check_r1(root, errors, verbose):
    """R1: Structural completeness."""
    # style must have class and version
    style_class = root.get("class")
    style_version = root.get("version")
    if not style_class:
        errors.append({"rule": "R1", "message": "<style> missing 'class' attribute", "severity": "error"})
    if not style_version:
        errors.append({"rule": "R1", "message": "<style> missing 'version' attribute", "severity": "error"})

    # info must have title, id, updated
    info = root.find("csl:info", NS)
    if info is None:
        errors.append({"rule": "R1", "message": "<info> element not found", "severity": "error"})
    else:
        for tag in ("title", "id", "updated"):
            el = info.find(f"csl:{tag}", NS)
            if el is None or not (el.text and el.text.strip()):
                errors.append({"rule": "R1", "message": f"<info> missing or empty <{tag}>", "severity": "error"})

    # citation must exist with layout
    citation = root.find("csl:citation", NS)
    if citation is None:
        errors.append({"rule": "R1", "message": "<citation> element not found", "severity": "error"})
    else:
        layouts = citation.findall("csl:layout", NS)
        if not layouts:
            errors.append({"rule": "R1", "message": "<citation> has no <layout> child", "severity": "error"})

    log_verbose(f"R1: {len([e for e in errors if e['rule'] == 'R1'])} issues", verbose)


def _check_r2(root, errors, verbose):
    """R2: Macro reference integrity."""
    # Collect defined macros
    defined = set()
    for macro in root.findall("csl:macro", NS):
        name = macro.get("name")
        if name:
            defined.add(name)

    # Collect referenced macros (text[@macro] anywhere in the tree)
    referenced = set()
    for el in root.iter(f"{{{CSL_NS}}}text"):
        macro_ref = el.get("macro")
        if macro_ref:
            referenced.add(macro_ref)

    # Also check <names> with <substitute> -> <names> pattern is ok, but
    # text[@macro] in any element
    for el in root.iter():
        macro_ref = el.get("macro")
        if macro_ref and el.tag == f"{{{CSL_NS}}}text":
            referenced.add(macro_ref)

    # Referenced but not defined -> error
    for name in sorted(referenced - defined):
        errors.append({
            "rule": "R2",
            "message": f"Macro '{name}' referenced but not defined",
            "severity": "error",
        })

    # Defined but never referenced -> warning
    for name in sorted(defined - referenced):
        errors.append({
            "rule": "R2",
            "message": f"Macro '{name}' defined but never referenced",
            "severity": "warning",
        })

    log_verbose(f"R2: defined={len(defined)}, referenced={len(referenced)}", verbose)


def _check_r3(root, errors, verbose):
    """R3: Class consistency — citation-format should match class."""
    style_class = root.get("class")
    if not style_class:
        return

    # Find citation-format from <category citation-format="..."/>
    citation_format = None
    for cat in root.iter(f"{{{CSL_NS}}}category"):
        cf = cat.get("citation-format")
        if cf:
            citation_format = cf
            break

    if citation_format is None:
        log_verbose("R3: No citation-format category found, skipping", verbose)
        return

    if style_class == "in-text":
        valid_formats = ("numeric", "author-date", "author", "label")
        if citation_format not in valid_formats:
            errors.append({
                "rule": "R3",
                "message": f"class='in-text' but citation-format='{citation_format}' "
                           f"(expected one of: {', '.join(valid_formats)})",
                "severity": "error",
            })
    elif style_class == "note":
        if citation_format != "note":
            errors.append({
                "rule": "R3",
                "message": f"class='note' but citation-format='{citation_format}' (expected 'note')",
                "severity": "error",
            })

    log_verbose(f"R3: class={style_class}, citation-format={citation_format}", verbose)


def _check_r4(root, errors, verbose):
    """R4: et-al-min > et-al-use-first for all <name> elements and inherited attributes."""
    count = 0

    # Gather elements that can carry et-al-min / et-al-use-first:
    # <name>, <style>, <citation>, <bibliography>
    targets = []
    for tag in ("name", "style", "citation", "bibliography"):
        for el in root.iter(f"{{{CSL_NS}}}{tag}"):
            targets.append((tag, el))

    for tag, el in targets:
        ea_min_str = el.get("et-al-min")
        ea_first_str = el.get("et-al-use-first")

        if ea_min_str is not None and ea_first_str is not None:
            try:
                ea_min = int(ea_min_str)
                ea_first = int(ea_first_str)
            except ValueError:
                errors.append({
                    "rule": "R4",
                    "message": f"<{tag}> has non-integer et-al-min='{ea_min_str}' or et-al-use-first='{ea_first_str}'",
                    "severity": "error",
                })
                continue

            if ea_min <= ea_first:
                errors.append({
                    "rule": "R4",
                    "message": f"<{tag}> has et-al-min={ea_min} <= et-al-use-first={ea_first} "
                               f"(et-al-min must be greater)",
                    "severity": "error",
                })
            count += 1

    log_verbose(f"R4: checked {count} elements with et-al attributes", verbose)


def _check_r5(root, errors, verbose):
    """R5: Bilingual layout ordering — layouts with locale attribute should come before those without."""
    for parent_tag in ("citation", "bibliography"):
        parent = root.find(f"csl:{parent_tag}", NS)
        if parent is None:
            continue

        layouts = parent.findall("csl:layout", NS)
        if len(layouts) <= 1:
            continue

        # Check ordering: locale-specific layouts should precede the generic one
        found_generic = False
        for layout in layouts:
            has_locale = layout.get("locale") is not None
            if not has_locale:
                found_generic = True
            elif found_generic:
                # A locale-specific layout appears after a generic one
                locale_val = layout.get("locale", "")
                errors.append({
                    "rule": "R5",
                    "message": f"In <{parent_tag}>: layout with locale='{locale_val}' "
                               f"appears after a layout without locale attribute "
                               f"(locale-specific layouts should come first)",
                    "severity": "warning",
                })

    log_verbose("R5: layout ordering checked", verbose)


def _check_r6(root, errors, verbose):
    """R6: No residual placeholders or empty macros."""
    # Check for placeholder text patterns like [占位符], [placeholder], [TODO], etc.
    placeholder_pattern = re.compile(r"\[.*?占位.*?\]|\[placeholder\]|\[TODO\]|\[FIXME\]|\[TBD\]", re.IGNORECASE)

    for el in root.iter():
        if not isinstance(el.tag, str):
            continue  # skip comments and processing instructions
        tag_local = etree.QName(el.tag).localname
        # Check text content
        if el.text:
            matches = placeholder_pattern.findall(el.text)
            for m in matches:
                errors.append({
                    "rule": "R6",
                    "message": f"Residual placeholder text found: '{m}' in <{tag_local}>",
                    "severity": "error",
                })
        # Check attribute values
        for attr_name, attr_val in el.attrib.items():
            matches = placeholder_pattern.findall(attr_val)
            for m in matches:
                errors.append({
                    "rule": "R6",
                    "message": f"Residual placeholder in attribute {attr_name}='{attr_val}' of <{tag_local}>",
                    "severity": "error",
                })
        # Check tail text
        if el.tail:
            matches = placeholder_pattern.findall(el.tail)
            for m in matches:
                errors.append({
                    "rule": "R6",
                    "message": f"Residual placeholder text found in tail: '{m}'",
                    "severity": "error",
                })

    # Check for empty macros (macros with no child elements)
    for macro in root.findall("csl:macro", NS):
        name = macro.get("name", "(unnamed)")
        if len(macro) == 0:
            # No child elements at all
            errors.append({
                "rule": "R6",
                "message": f"Empty macro '{name}' (no child elements)",
                "severity": "error",
            })

    log_verbose("R6: placeholder and empty macro check done", verbose)


# ---------------------------------------------------------------------------
# Main
# ---------------------------------------------------------------------------

def validate_csl(filepath, verbose=False):
    """Run all validation stages and return the result dict."""
    filepath = str(Path(filepath).resolve())

    if not os.path.isfile(filepath):
        return {
            "file": filepath,
            "stages": [],
            "overall": "FAIL",
            "error": f"File not found: {filepath}",
        }

    stages = []

    # Stage 1
    if verbose:
        print("Stage 1: XML Syntax Check...", file=sys.stderr)
    s1_result, tree = stage1_xml_syntax(filepath, verbose)
    stages.append(s1_result)

    # Stage 2
    if verbose:
        print("Stage 2: CSL Schema Validation...", file=sys.stderr)
    s2_result = stage2_schema_validation(tree, verbose)
    stages.append(s2_result)

    # Stage 3
    if verbose:
        print("Stage 3: Logic Rules Check...", file=sys.stderr)
    s3_result = stage3_logic_rules(tree, verbose)
    stages.append(s3_result)

    # Determine overall result
    all_errors = []
    for s in stages:
        all_errors.extend(s.get("errors", []))

    has_error = any(e["severity"] == "error" for e in all_errors)  # noqa: F841
    has_warning = any(e["severity"] == "warning" for e in all_errors)  # noqa: F841

    # Skipped stages don't count as errors for overall if they only have warnings
    skipped_only_warnings = True
    for s in stages:
        if s.get("skipped"):
            if any(e["severity"] == "error" for e in s.get("errors", [])):
                skipped_only_warnings = False  # noqa: F841

    # Recalculate has_error excluding skipped-stage warnings
    real_errors = []
    for s in stages:
        if s.get("skipped"):
            continue
        real_errors.extend(e for e in s.get("errors", []) if e["severity"] == "error")

    real_warnings = []
    for s in stages:
        if s.get("skipped"):
            # Skipped stage warnings are informational
            real_warnings.extend(e for e in s.get("errors", []) if e["severity"] == "warning")
        else:
            real_warnings.extend(e for e in s.get("errors", []) if e["severity"] == "warning")

    if real_errors:
        overall = "FAIL"
    elif real_warnings:
        overall = "WARN"
    else:
        overall = "PASS"

    return {
        "file": filepath,
        "stages": stages,
        "overall": overall,
    }


def main():
    parser = argparse.ArgumentParser(
        description="Validate a CSL (Citation Style Language) file.",
        epilog="Example: python validate_csl.py --verbose style.csl",
    )
    parser.add_argument("file", help="Path to the CSL file to validate")
    parser.add_argument("--verbose", "-v", action="store_true",
                        help="Show detailed progress on stderr")
    args = parser.parse_args()

    result = validate_csl(args.file, verbose=args.verbose)
    print(json.dumps(result, ensure_ascii=False, indent=2))

    # Exit code
    if result["overall"] == "FAIL":
        sys.exit(1)
    elif result["overall"] == "WARN":
        sys.exit(0)
    else:
        sys.exit(0)


if __name__ == "__main__":
    main()