Merge pull request #496 from xxyzz/fr

Update fr edition extractor

src/wiktextract/data/fr/linkage_subtitles.json

@@ -1,6 +1,7 @@
 {
   "abrév": "abbreviation",
   "abréviations": "abbreviation",
+  "antonymes": "antonyms",
   "app": "related",
   "apparentés": "related",
   "apr": "related",

src/wiktextract/extractor/fr/gloss.py

@@ -70,6 +70,8 @@ def extract_gloss(
         find_alt_of_form(
             wxr, gloss_only_nodes[:note_index], page_data[-1].pos, gloss_data
         )
+        if "form-of" in page_data[-1].tags:
+            find_form_of_word(wxr, gloss_only_nodes[:note_index], gloss_data)
         gloss_text = clean_node(
             wxr, gloss_data, gloss_only_nodes[:note_index]
         ).strip(" ()")
@@ -192,3 +194,16 @@ def find_alt_of_form(
                 alt_of = clean_node(wxr, None, link)
         if len(alt_of) > 0:
             gloss_data.alt_of.append(AltForm(word=alt_of))
+
+
+def find_form_of_word(
+    wxr: WiktextractContext,
+    gloss_nodes: list[Union[str, WikiNode]],
+    gloss_data: Sense,
+) -> None:
+    form_of = ""
+    for node in gloss_nodes:
+        if isinstance(node, WikiNode) and node.kind == NodeKind.LINK:
+            form_of = clean_node(wxr, None, node)
+    if len(form_of) > 0:
+        gloss_data.form_of.append(AltForm(word=form_of))

src/wiktextract/extractor/fr/inflection.py

@@ -31,6 +31,9 @@ def extract_inflection(
         "commun",  # sv-nom-c-ar
     }
 )
+IGNORE_TABLE_HEADER_PREFIXES = (
+    "voir la conjugaison du verbe",  # Modèle:fr-verbe-flexion
+)
 IGNORE_TABLE_CELL = frozenset(
     {
         "Déclinaisons",  # de-adj
@@ -108,7 +111,13 @@ def process_inflection_table(
                     table_header_text = clean_node(
                         wxr, None, table_cell
                     ).replace("\n", " ")
-                    if table_header_text.lower() in IGNORE_TABLE_HEADERS:
+                    if (
+                        table_header_text.lower() in IGNORE_TABLE_HEADERS
+                        or table_header_text.lower().startswith(
+                            IGNORE_TABLE_HEADER_PREFIXES
+                        )
+                        or len(table_header_text.strip()) == 0
+                    ):
                         continue
                     if not current_row_has_data_cell:
                         # if all cells of the row are header cells

src/wiktextract/extractor/fr/models.py

@@ -87,6 +87,7 @@ class Sense(FrenchBaseModel):
     examples: list[Example] = []
     note: str = ""
     alt_of: list[AltForm] = []
+    form_of: list[AltForm] = []
 
 
 class WordEntry(FrenchBaseModel):
@@ -104,6 +105,7 @@ class WordEntry(FrenchBaseModel):
     forms: list[Form] = Field([], description="Inflection forms list")
     sounds: list[Sound] = []
     translations: list[Translation] = []
+    antonyms: list[Linkage] = []
     synonyms: list[Linkage] = []
     hyponyms: list[Linkage] = []
     hypernyms: list[Linkage] = []

src/wiktextract/extractor/fr/page.py

@@ -2,7 +2,7 @@
 from typing import Any, Optional
 
 from wikitextprocessor import NodeKind, WikiNode
-from wikitextprocessor.parser import LEVEL_KIND_FLAGS, TemplateNode
+from wikitextprocessor.parser import LEVEL_KIND_FLAGS
 from wiktextract.page import clean_node
 from wiktextract.wxr_context import WiktextractContext
 
@@ -109,7 +109,7 @@ def process_pos_block(
     wxr: WiktextractContext,
     page_data: list[WordEntry],
     base_data: WordEntry,
-    pos_title_node: TemplateNode,
+    pos_title_node: WikiNode,
     pos_argument: str,
     pos_title: str,
 ):
@@ -120,6 +120,12 @@ def process_pos_block(
     page_data[-1].pos = pos_type
     page_data[-1].pos_title = pos_title
     page_data[-1].tags.extend(pos_data.get("tags", []))
+    for level_node_template in pos_title_node.find_content(NodeKind.TEMPLATE):
+        if (
+            level_node_template.template_name == "S"
+            and level_node_template.template_parameters.get(3) == "flexion"
+        ):
+            page_data[-1].tags.append("form-of")
     child_nodes = list(pos_title_node.filter_empty_str_child())
     form_line_start = 0  # Ligne de forme
     gloss_start = len(child_nodes)

src/wiktextract/extractor/fr/translation.py

@@ -64,7 +64,9 @@ def process_italic_node(
         and previous_node.template_name.startswith("trad")
         and len(page_data[-1].translations) > 0
     ):
-        page_data[-1].translations[-1].tags.append(tag.strip("()"))
+        tag = tag.strip("()")
+        if len(tag) > 0:
+            page_data[-1].translations[-1].tags.append(tag)
 
 
 def process_translation_templates(
@@ -143,4 +145,5 @@ def process_translation_templates(
         page_data[-1].translations.append(translation_data)
     elif len(page_data[-1].translations) > 0:
         tag = clean_node(wxr, None, template_node).strip("()")
-        page_data[-1].translations[-1].tags.append(tag)
+        if len(tag) > 0:
+            page_data[-1].translations[-1].tags.append(tag)

src/wiktextract/extractor/share.py

@@ -29,7 +29,9 @@ def capture_text_in_parentheses(text: str) -> tuple[list[str], str]:
         if len(not_captured) > 0:
             rest_parts.append(not_captured)
         last_group_end = m.end()
-        capture_text_list.append(m.group()[1:-1])
+        text = m.group()[1:-1].strip()
+        if len(text) > 0:
+            capture_text_list.append(text)
 
     rest_text = " ".join(rest_parts) if len(rest_parts) > 0 else text
     return capture_text_list, rest_text

tests/test_fr_gloss.py

@@ -5,14 +5,20 @@
 from wiktextract.config import WiktionaryConfig
 from wiktextract.extractor.fr.gloss import extract_gloss
 from wiktextract.extractor.fr.models import WordEntry
-from wiktextract.extractor.fr.page import process_pos_block
+from wiktextract.extractor.fr.page import parse_page, process_pos_block
 from wiktextract.wxr_context import WiktextractContext
 
 
 class TestFrGloss(TestCase):
+    maxDiff = None
+
     def setUp(self) -> None:
         self.wxr = WiktextractContext(
-            Wtp(lang_code="fr"), WiktionaryConfig(dump_file_lang_code="fr")
+            Wtp(lang_code="fr"),
+            WiktionaryConfig(
+                dump_file_lang_code="fr",
+                capture_language_codes=None,
+            ),
         )
 
     def tearDown(self) -> None:
@@ -392,3 +398,42 @@ def test_variante_de_dif(self):
                 }
             ],
         )
+
+    def test_form_of(self):
+        self.wxr.wtp.start_page("dièse")
+        self.wxr.wtp.add_page("Modèle:langue", 10, "Français")
+        self.wxr.wtp.add_page("Modèle:S", 10, "Forme de verbe")
+        self.assertEqual(
+            parse_page(
+                self.wxr,
+                "dièse",
+                """== {{langue|fr}} ==
+=== {{S|verbe|fr|flexion}} ===
+# ''Première personne du singulier de l’indicatif présent du verbe'' [[diéser]].
+# ''Troisième personne du singulier de l’indicatif présent du verbe'' [[diéser]].""",
+            ),
+            [
+                {
+                    "lang": "Français",
+                    "lang_code": "fr",
+                    "pos": "verb",
+                    "pos_title": "Forme de verbe",
+                    "senses": [
+                        {
+                            "form_of": [{"word": "diéser"}],
+                            "glosses": [
+                                "Première personne du singulier de l’indicatif présent du verbe diéser."
+                            ],
+                        },
+                        {
+                            "form_of": [{"word": "diéser"}],
+                            "glosses": [
+                                "Troisième personne du singulier de l’indicatif présent du verbe diéser."
+                            ],
+                        },
+                    ],
+                    "tags": ["form-of"],
+                    "word": "dièse"
+                }
+            ],
+        )

tests/test_fr_linkage.py

@@ -230,3 +230,18 @@ def test_italic_sense_node(self):
                 {"word": "more", "sense": "selon les adjectifs"},
             ],
         )
+
+    def test_no_linkage_empty_tag(self):
+        page_data = [WordEntry(word="gambo", lang_code="eo", lang="Espéranto")]
+        self.wxr.wtp.start_page("gambo")
+        root = self.wxr.wtp.parse("* [[korpo]] ( ''[[corps]]'' )")
+        extract_linkage(self.wxr, page_data, root, "holonymes")
+        self.assertEqual(
+            [
+                d.model_dump(exclude_defaults=True)
+                for d in page_data[-1].holonyms
+            ],
+            [
+                {"word": "korpo", "sense": "corps"},
+            ],
+        )

tests/test_fr_page.py

@@ -11,13 +11,16 @@
 
 
 class TestFrPage(TestCase):
+    maxDiff = None
+
     def setUp(self):
-        self.maxDiff = None
-        conf1 = WiktionaryConfig(
-            dump_file_lang_code="fr",
-            capture_language_codes=None,
+        self.wxr = WiktextractContext(
+            Wtp(lang_code="fr"),
+            WiktionaryConfig(
+                dump_file_lang_code="fr",
+                capture_language_codes=None,
+            ),
         )
-        self.wxr = WiktextractContext(Wtp(lang_code="fr"), conf1)
 
     def tearDown(self) -> None:
         self.wxr.wtp.close_db_conn()

tests/test_fr_pronunciation.py

@@ -111,7 +111,7 @@ def test_no_ipa(self):
         self.wxr.wtp.add_page(
             "Modèle:écouter",
             10,
-            '<span><span>Suède</span>&nbsp;: écouter «&nbsp;<span>mars</span> <span><span>[<small><span>[//fr.wiktionary.org Prononciation ?]</span></small>]</span></span>&nbsp;» <span>[[File:LL-Q9027 (swe)-Moonhouse-mars.wav]]</span></span>',
+            "<span><span>Suède</span>&nbsp;: écouter «&nbsp;<span>mars</span> <span><span>[<small><span>[//fr.wiktionary.org Prononciation ?]</span></small>]</span></span>&nbsp;» <span>[[File:LL-Q9027 (swe)-Moonhouse-mars.wav]]</span></span>",
         )
         extract_pronunciation(
             self.wxr,