biblatex-csl-converter/src/export/csl/index.ts

a set of converters: biblatex => json, CSL => json, json => biblatex, json => CSL

import {
    BibTypes,
    BibFieldTypes,
    NodeArray,
    NameDictObject,
    NodeObject,
} from "../../const"
import { edtfParse } from "../../edtf-parser"
import { BibDB } from "../../import/biblatex"
import { toSentenceCase } from "./sentence-caser"

/** Converts a BibDB to a DB of the CSL type.
 * @param bibDB The bibliography database to convert.
 */
export interface TagEntry {
    open: string
    close: string
}

export type Tags = Record<string, TagEntry>

const TAGS: Tags = {
    strong: { open: "<b>", close: "</b>" },
    em: { open: "<i>", close: "</i>" },
    sub: { open: "<sub>", close: "</sub>" },
    sup: { open: "<sup>", close: "</sup>" },
    smallcaps: {
        open: '<span style="font-variant:small-caps;">',
        close: "</span>",
    },
    nocase: { open: '<span class="nocase">', close: "</span>" },
    enquote: { open: "“", close: "”" },
    url: { open: "", close: "" },
    undefined: { open: "[", close: "]" },
}

type ConfigObject = {
    escapeText?: boolean
    useEntryKeys?: boolean // Whether to output using the entry keys
    language?: string // Language of the citation
}

type ErrorObject = {
    type: string
    variable: string
}

type CSLDateObject = {
    "date-parts"?: [number[]] | [number[], number[]]
    circa?: boolean
}

type CSLNameObject = {
    literal?: string
    given?: string
    family?: string
    suffix?: string
    "non-dropping-particle"?: string
    "dropping-particle"?: string
}

export interface CSLEntry {
    id?: string
    [key: string]: unknown
}

export type CSLOutput = Record<string, CSLEntry>

export class CSLExporter {
    bibDB: BibDB
    pks: string[]
    config: ConfigObject
    cslDB: Record<string, CSLEntry>
    errors: ErrorObject[]

    constructor(
        bibDB: BibDB,
        pks: string[] | false = false,
        config: ConfigObject = {}
    ) {
        this.bibDB = bibDB
        if (pks) {
            this.pks = pks // A list of pk values of the bibliography items to be exported.
        } else {
            this.pks = Object.keys(bibDB) // If none are selected, all keys are exported.
        }
        this.config = config
        this.cslDB = {}
        this.errors = []
    }

    parse(): CSLOutput {
        for (let bibId in this.bibDB) {
            if (this.pks.indexOf(bibId) !== -1) {
                const id = this.config.useEntryKeys
                    ? this.bibDB[bibId].entry_key || bibId
                    : bibId
                this.cslDB[id] = this.getCSLEntry(bibId)
                this.cslDB[id].id = id
            }
        }
        return this.cslDB
    }
    /** Converts one BibDB entry to CSL format.
     * @function getCSLEntry
     * @param id The id identifying the bibliography entry.
     */
    getCSLEntry(id: string): CSLEntry {
        let bib = this.bibDB[id as unknown as number],
            fValues: CSLEntry = {}
        if (!bib.fields || !bib.bib_type || !BibTypes[bib.bib_type]) {
            return fValues
        }

        // If there's a language field, use it for deciding whether to apply sentence casing
        if (bib.fields.langid) {
            this.config.language = bib.fields.langid as string
        }

        for (let fKey in bib.fields) {
            if (
                bib.fields[fKey] !== "" &&
                fKey in BibFieldTypes &&
                "csl" in BibFieldTypes[fKey]
            ) {
                let fValue = bib.fields[fKey]
                let fType = BibFieldTypes[fKey]["type"]
                let key: string
                const csl = BibFieldTypes[fKey].csl!
                if (typeof csl === "string") {
                    key = csl
                } else if (csl[bib.bib_type]) {
                    key = csl[bib.bib_type]
                } else {
                    key = csl["*"]
                }
                let reformedValue
                switch (fType) {
                    case "f_date":
                        reformedValue = this._reformDate(fValue as string)
                        if (reformedValue) {
                            fValues[key] = reformedValue
                        }
                        break
                    case "f_integer":
                        fValues[key] = this._reformInteger(fValue)
                        break
                    case "f_key":
                        fValues[key] = this._reformKey(fValue, fKey)
                        break
                    case "f_literal":
                    case "f_long_literal":
                        fValues[key] = this._reformText(fValue)
                        break
                    case "l_range":
                        fValues[key] = this._reformRange(fValue)
                        break
                    case "f_title":
                        fValues[key] = this._reformTitle(fValue)
                        break
                    case "f_uri":
                    case "f_verbatim":
                        fValues[key] = fValue
                        break
                    case "l_key":
                        fValues[key] = (fValue as (string | NodeArray)[])
                            .map((key: string | NodeArray) => {
                                return this._reformKey(key, fKey)
                            })
                            .join(" and ")
                        break
                    case "l_literal":
                        fValues[key] = (fValue as NodeArray[])
                            .map((text: NodeArray) => this._reformText(text))
                            .join(", ")
                        break
                    case "l_name":
                        fValues[key] = this._reformName(
                            fValue as NameDictObject[]
                        )
                        break
                    case "l_tag":
                        fValues[key] = (fValue as string[]).join(", ")
                        break
                    default:
                        console.warn(`Unrecognized field type: ${fType}!`)
                }
            }
        }
        fValues["type"] = BibTypes[bib.bib_type].csl
        return fValues
    }

    _reformKey(theValue: string | unknown, fKey: string): string {
        if (typeof theValue === "string") {
            let fieldType = BibFieldTypes[fKey]!
            if (Array.isArray(fieldType["options"])) {
                return theValue
            } else {
                return fieldType["options"]![theValue]["csl"]
            }
        } else {
            return this._reformText(theValue)
        }
    }

    _reformRange(theValue: unknown): string {
        if (!Array.isArray(theValue)) {
            console.warn(`Wrong format for range`, theValue)
            return ""
        }
        return theValue
            .map((interval) => this._reformInterval(interval))
            .filter((interval) => interval.length)
            .join(",")
    }

    _reformInterval(theValue: unknown): string {
        if (!Array.isArray(theValue)) {
            console.warn(`Wrong format for interval`, theValue)
            return ""
        }
        return theValue.map((text) => this._reformText(text)).join("-")
    }

    _reformInteger(theValue: unknown): string | number {
        let theString = this._reformText(theValue)
        let theInt = parseInt(theString)
        if (theString !== String(theInt)) {
            return theString
        }
        return theInt
    }

    _escapeText(theValue: unknown): string {
        if (!(typeof theValue === "string")) {
            console.warn(`Wrong format for escapeText`, theValue)
            return ""
        }
        return theValue
            .replace(/&/g, "&amp;")
            .replace(/</g, "&lt;")
            .replace(/>/g, "&gt;")
            .replace(/'/g, "&apos;")
            .replace(/"/g, "&quot;")
    }

    _reformTitle(theValue: unknown): string {
        let html = "",
            lastMarks: string[] = []
        if (!Array.isArray(theValue)) {
            console.warn(`Wrong format for reformTitle`, theValue)
            return html
        }

        // Get the plain text to potentially apply sentence casing
        const plainText = theValue.reduce((text, node) => {
            if ("text" in node) {
                return text + node.text
            } else {
                return text
            }
        }, "")

        // Check if the language is English or not specified
        // Check against all English language variants based on langidOptions values
        // Only apply sentence casing for English (default) language
        const isEnglishLanguage =
            !this.config.language ||
            this.config.language.toLowerCase().endsWith("english")

        // Only apply sentence casing for English or unspecified language
        let sentenceCasedText: string = isEnglishLanguage
            ? toSentenceCase(plainText)
            : plainText

        theValue.forEach((node: NodeObject) => {
            if (node.type === "variable") {
                // This is an undefined variable
                // This should usually not happen, as CSL doesn't know what to
                // do with these. We'll put them into an unsupported tag.
                html += `${TAGS.undefined.open}${node.attrs!.variable}${
                    TAGS.undefined.close
                }`
                this.errors.push({
                    type: "undefined_variable",
                    variable: node.attrs!.variable as string,
                })
                return
            }
            let newMarks = node.marks ? node.marks.map((mark) => mark.type) : []
            // close all tags that are not present in current text node.
            let closing = false,
                closeTags: string[] = []
            lastMarks.forEach((mark, index) => {
                if (mark != newMarks[index]) {
                    closing = true
                }
                if (closing) {
                    closeTags.push(TAGS[mark].close)
                }
            })
            // Add close tags in reverse order to close innermost tags
            // first.
            closeTags.reverse()
            html += closeTags.join("")

            // open all new tags that were not present in the last text node.
            let opening = false
            newMarks.forEach((mark, index) => {
                if (mark != lastMarks[index]) {
                    opening = true
                }
                if (opening) {
                    html += TAGS[mark].open
                }
            })
            if ("text" in node) {
                const sentencesCasedNodeText = sentenceCasedText.substr(
                    0,
                    node.text.length
                )
                sentenceCasedText = sentenceCasedText.slice(node.text.length)

                const insertedText = newMarks.find((mark) => mark === "nocase")
                    ? node.text
                    : sentencesCasedNodeText

                html += this.config.escapeText
                    ? this._escapeText(insertedText)
                    : insertedText
            }
            lastMarks = newMarks
        })
        // Close all still open tags
        lastMarks
            .slice()
            .reverse()
            .forEach((mark) => {
                html += TAGS[mark].close
            })
        return html
    }

    _reformText(theValue: unknown): string {
        let html = "",
            lastMarks: string[] = []
        if (!Array.isArray(theValue)) {
            console.warn(`Wrong format for reformText`, theValue)
            return html
        }
        theValue.forEach((node: NodeObject) => {
            if (node.type === "variable") {
                // This is an undefined variable
                // This should usually not happen, as CSL doesn't know what to
                // do with these. We'll put them into an unsupported tag.
                html += `${TAGS.undefined.open}${node.attrs!.variable}${
                    TAGS.undefined.close
                }`
                this.errors.push({
                    type: "undefined_variable",
                    variable: node.attrs!.variable as string,
                })
                return
            }
            let newMarks = node.marks ? node.marks.map((mark) => mark.type) : []
            // close all tags that are not present in current text node.
            let closing = false,
                closeTags: string[] = []
            lastMarks.forEach((mark, index) => {
                if (mark != newMarks[index]) {
                    closing = true
                }
                if (closing) {
                    closeTags.push(TAGS[mark].close)
                }
            })
            // Add close tags in reverse order to close innermost tags
            // first.
            closeTags.reverse()
            html += closeTags.join("")

            // open all new tags that were not present in the last text node.
            let opening = false
            newMarks.forEach((mark, index) => {
                if (mark != lastMarks[index]) {
                    opening = true
                }
                if (opening) {
                    html += TAGS[mark].open
                }
            })
            if ("text" in node) {
                html += this.config.escapeText
                    ? this._escapeText(node.text)
                    : node.text
            }
            lastMarks = newMarks
        })
        // Close all still open tags
        lastMarks
            .slice()
            .reverse()
            .forEach((mark) => {
                html += TAGS[mark].close
            })
        return html
    }

    _reformDate(dateStr: string): false | CSLDateObject {
        let dateObj = edtfParse(dateStr)
        const reformedDate: CSLDateObject = {}
        if (!dateObj.valid) {
            return false
        } else if (
            dateObj.values.length > 1 &&
            Array.isArray(dateObj.values[0]) &&
            Array.isArray(dateObj.values[1])
        ) {
            if (
                dateObj.values[0][0] === 0 &&
                !String(dateObj.values[1]).replace(/9/g, "").length
            ) {
                // both from and to year are completely unknown.
                return false
            }
            const intervalFrom: Array<number | string> = dateObj.values[0],
                intervalTo: Array<number | string> = dateObj.values[1]
            const intervalDateParts: [number[], number[]] = [
                intervalFrom
                    .slice(0, 3)
                    .map((value) => parseInt(value as string)),
                intervalTo
                    .slice(0, 3)
                    .map((value) => parseInt(value as string)),
            ]
            reformedDate["date-parts"] = intervalDateParts
        } else {
            const values: number[] = dateObj.values
                .slice(0, 3)
                .map((value) => parseInt(value as string))
            reformedDate["date-parts"] = [values]
            if (dateObj.type === "Interval") {
                // Open interval that we cannot represent, so we make it circa instead.
                reformedDate["circa"] = true
            }
        }

        if (dateObj.uncertain || dateObj.approximate) {
            reformedDate["circa"] = true
        }
        return reformedDate
    }

    _reformName(theNames: NameDictObject[]): CSLNameObject[] {
        const names = theNames.map((name) => {
            const reformedName: CSLNameObject = {}
            if (name.literal) {
                let literal = this._reformText(name.literal)
                if (literal.length) {
                    reformedName["literal"] = literal
                } else {
                    return false
                }
            } else {
                reformedName["given"] = this._reformText(name.given!)
                reformedName["family"] = this._reformText(name.family!)
                if (name.suffix) {
                    reformedName["suffix"] = this._reformText(name.suffix)
                }
                if (name.prefix) {
                    if (name.useprefix) {
                        reformedName["non-dropping-particle"] =
                            this._reformText(name.prefix)
                    } else {
                        reformedName["dropping-particle"] = this._reformText(
                            name.prefix
                        )
                    }
                }
                reformedName["family"] = this._reformText(name["family"]!)
            }
            return reformedName
        })
        return names.filter((name) => name) as CSLNameObject[]
    }
}