[de-en] fix missing propositions (#127)

yomidevs · Aug 7, 2024 · 097927c · 097927c
1 parent 53a8943
commit 097927c
Show file tree

Hide file tree

Showing 6 changed files with 97 additions and 3 deletions.
diff --git a/3-tidy-up.js b/3-tidy-up.js
@@ -353,11 +353,16 @@ function getCanonicalForm({word, forms}) {
         form.tags &&
         form.tags.includes('canonical')
     );
-    if (canonicalForm) {
+    if (canonicalForm && canonicalForm.form) {
         word = canonicalForm.form;
 
-        if (word && word.includes('{{#ifexist:Wiktionary')) { // TODO: remove once fixed in kaikki
-            word = word.replace(/ {{#if:.+/, '');
+        if (word.includes('{{#ifexist:Wiktionary')) { // TODO: remove once fixed in kaikki
+            word = word.replace(/ {{#if:.+/, '').trim();
+        }
+
+        bracketsRegex = /\[.*\]$/;
+        if (bracketsRegex.test(word)) {
+            word = word.replace(bracketsRegex, '').trim();
         }
     }
     return word;

diff --git a/data/test/dict/de/en/tag_bank_1.json b/data/test/dict/de/en/tag_bank_1.json
@@ -96,5 +96,12 @@
     1,
     "rare",
     -1
+  ],
+  [
+    "prep",
+    "partOfSpeech",
+    -1,
+    "preposition",
+    1
   ]
 ]
diff --git a/data/test/dict/de/en/term_bank_1.json b/data/test/dict/de/en/term_bank_1.json
@@ -351,5 +351,21 @@
     ],
     0,
     ""
+  ],
+  [
+    "von",
+    "",
+    "prep",
+    "prep",
+    0,
+    [
+      "from",
+      "of, belonging to (often replacing genitive; see usage note below)",
+      "by (with passive voice)",
+      "about, of (a topic)",
+      "on, with (a resource)"
+    ],
+    0,
+    ""
   ]
 ]
diff --git a/data/test/ipa/de/en/term_meta_bank_1.json b/data/test/ipa/de/en/term_meta_bank_1.json
@@ -70,5 +70,22 @@
         }
       ]
     }
+  ],
+  [
+    "von",
+    "ipa",
+    {
+      "reading": "von",
+      "transcriptions": [
+        {
+          "ipa": "/fɔn/",
+          "tags": []
+        },
+        {
+          "ipa": "[fɔn]",
+          "tags": []
+        }
+      ]
+    }
   ]
 ]
diff --git a/data/test/kaikki/de-en.json b/data/test/kaikki/de-en.json
diff --git a/data/test/tidy/de-en-lemmas.json b/data/test/tidy/de-en-lemmas.json
@@ -442,5 +442,53 @@
         ]
       }
     }
+  },
+  "von": {
+    "von": {
+      "prep": {
+        "ipa": [
+          {
+            "ipa": "/fɔn/",
+            "tags": []
+          },
+          {
+            "ipa": "[fɔn]",
+            "tags": []
+          }
+        ],
+        "senses": [
+          {
+            "glosses": [
+              "from"
+            ],
+            "tags": []
+          },
+          {
+            "glosses": [
+              "of, belonging to (often replacing genitive; see usage note below)"
+            ],
+            "tags": []
+          },
+          {
+            "glosses": [
+              "by (with passive voice)"
+            ],
+            "tags": []
+          },
+          {
+            "glosses": [
+              "about, of (a topic)"
+            ],
+            "tags": []
+          },
+          {
+            "glosses": [
+              "on, with (a resource)"
+            ],
+            "tags": []
+          }
+        ]
+      }
+    }
   }
 }
-Original file line number
+Diff line change
@@ Expand Up / @@ -96,5 +96,12 @@ @@
 ,
         "rare",
         -1
+      ],
+      [
+        "prep",
+        "partOfSpeech",
+        -1,
+        "preposition",
       ]
     ]