"""Documentation links for the UPOS and basic dependency labels in a parse."""
import html

POS_NAMES = {
    "ADJ": "Adjective", "ADP": "Adposition", "ADV": "Adverb",
    "AUX": "Auxiliary", "CCONJ": "Coordinating conjunction", "DET": "Determiner",
    "INTJ": "Interjection", "NOUN": "Noun", "NUM": "Numeral", "PART": "Particle",
    "PRON": "Pronoun", "PROPN": "Proper noun", "PUNCT": "Punctuation",
    "SCONJ": "Subordinating conjunction", "SYM": "Symbol", "VERB": "Verb", "X": "Other",
}
RELATION_NAMES = {
    "nsubj": "Nominal subject", "obj": "Object", "iobj": "Indirect object",
    "csubj": "Clausal subject", "ccomp": "Clausal complement",
    "xcomp": "Open clausal complement", "obl": "Oblique nominal",
    "vocative": "Vocative", "expl": "Expletive", "dislocated": "Dislocated element",
    "advcl": "Adverbial clause modifier", "advmod": "Adverbial modifier",
    "discourse": "Discourse element", "nmod": "Nominal modifier",
    "appos": "Appositional modifier", "nummod": "Numeric modifier",
    "acl": "Clausal modifier of a noun", "amod": "Adjectival modifier",
    "det": "Determiner", "clf": "Classifier", "case": "Case marking",
    "conj": "Conjunct", "cc": "Coordinating conjunction",
    "fixed": "Fixed multiword expression", "flat": "Flat multiword expression",
    "compound": "Compound", "list": "List", "parataxis": "Parataxis",
    "orphan": "Orphan", "goeswith": "Goes with", "reparandum": "Overridden disfluency",
    "punct": "Punctuation", "root": "Root", "dep": "Unspecified dependency",
    "aux": "Auxiliary", "cop": "Copula", "mark": "Marker",
}
# Verified universal subtype pages. Other subtypes link to their base relation
# rather than assuming that a universal page exists for a language-specific label.
SUBTYPE_NAMES = {
    "acl:relcl": "Relative clause modifier", "aux:pass": "Passive auxiliary",
    "csubj:pass": "Clausal passive subject", "nsubj:pass": "Passive nominal subject",
    "nmod:poss": "Possessive nominal modifier",
}

# Curated Universal Dependencies morphological features for English and German.
FEATURE_NAMES = {
    # 1. Core morphological features
    "Case": "Case (e.g. nominative, accusative, dative, genitive)",
    "Gender": "Gender (masculine, feminine, neuter)",
    "Number": "Number (singular, plural)",
    "Person": "Person (first, second, third)",
    "Tense": "Tense (past, present, future)",
    "Mood": "Mood (indicative, imperative, subjunctive)",
    "VerbForm": "Verb form (finite, infinitive, participle, gerund)",
    "Degree": "Degree of comparison (positive, comparative, superlative)",
    # 2. Additional relevant features
    "PronType": "Pronominal type (personal, relative, demonstrative, etc.)",
    "Definite": "Definiteness (definite, indefinite)",
    "Poss": "Possessive marker/determiner",
    "Reflex": "Reflexive pronoun/marker",
    "NumType": "Numeral type (cardinal, ordinal, etc.)",
    "Voice": "Voice (active, passive)",
    "Aspect": "Aspect (aspectual distinctions)",
    "Polarity": "Polarity (positive, negative)",
    # 3. Other features worth knowing
    "Abbr": "Abbreviation",
    "Foreign": "Foreign word",
    "Typo": "Typographical error / unnormalized spelling",
    "ExtPos": "External part of speech",
}


def render_ud_help(conllu_text, word_links_html=""):
    """Render expanded Help with deduplicated labels used by visible word rows."""
    pos, relations, features = set(), set(), set()
    for line in conllu_text.splitlines():
        fields = line.split("\t")
        # The renderer excludes multiword headers and empty nodes by default.
        if len(fields) != 10 or not fields[0].isdigit():
            continue
        if fields[3] in POS_NAMES:
            pos.add(fields[3])
        if fields[7].split(":")[0] in RELATION_NAMES:
            relations.add(fields[7])
        # FEATS is column 5 (0-indexed)
        feats_val = fields[5]
        if feats_val and feats_val != "_":
            for feat_pair in feats_val.split("|"):
                if "=" in feat_pair:
                    feat_name = feat_pair.split("=")[0].strip()
                    if feat_name in FEATURE_NAMES:
                        features.add(feat_name)

    def link(label, name, path):
        return (
            f'<li><a href="https://universaldependencies.org/u/{path}.html" '
            f'target="_blank" rel="noopener noreferrer">{html.escape(label)}</a>'
            f' — {html.escape(name)}</li>'
        )

    tags = "".join(link(tag, POS_NAMES[tag], "pos/" + tag) for tag in sorted(pos))
    dependencies = []
    for relation in sorted(relations):
        base = relation.split(":")[0]
        if relation in SUBTYPE_NAMES:
            name, page = SUBTYPE_NAMES[relation], relation.replace(":", "-")
        else:
            name, page = RELATION_NAMES[base], base
            if ":" in relation:
                name += " (subtype; link explains the base relation)"
        dependencies.append(link(relation, name, "dep/" + page))

    feat_links = "".join(
        link(feat, FEATURE_NAMES[feat], "feat/" + feat) for feat in sorted(features)
    )

    if not tags and not dependencies and not feat_links:
        return ""

    columns_html = ""
    if tags:
        columns_html += '<div><strong>POS tags in this sentence</strong><ul>' + tags + '</ul></div>'
    if dependencies:
        columns_html += '<div><strong>Relations in this sentence</strong><ul>' + "".join(dependencies) + '</ul></div>'
    if feat_links:
        columns_html += '<div><strong>Morphological features in this sentence</strong><ul>' + feat_links + '</ul></div>'
    if word_links_html:
        columns_html += word_links_html

    return (
        '<details class="ud-help" open><summary>Help</summary>'
        f'<div class="ud-help-columns">{columns_html}</div></details>'
    )