Phraseandspeed

Pipfile

@@ -11,7 +11,6 @@ pytest-cov = "*"
 pytest-datadir = "*"
 pytest-mypy = "*"
 hypothesis = {version = "~=4.34", extras = ["django"]}
-codecov = "*"
 pysnooper = "*"
 python-levenshtein = "*"
 django-debug-toolbar = "*"
@@ -21,7 +20,6 @@ mypy = "*"
 pytest-env = "*"
 jupyterlab = "*"
 appnope = "*"
-nb_black = "*"
 statsmodels = "*"
 pandas-stubs = "*"
 pytest-pythonpath = "*"

src/API/search/__init__.py

@@ -6,8 +6,8 @@ def search_with_affixes(
     rw_index: str,
     rw_domain: str,
     wn_synset: str,
-    include_auto_definitions=False,
-    inflect_english_phrases=False,
+    include_auto_definitions=True,
+    inflect_english_phrases=True,
 ):
     """
     Search for wordforms matching:

src/API/search/cvd_search.py

@@ -31,6 +31,7 @@ def do_cvd_search(search_run: SearchRun):
 
     search_run.add_verbose_message(cvd_extracted_keys=keys)
     query_vector = vector_for_keys(google_news_vectors(), keys)
+    print(query_vector)
 
     try:
         closest = definition_vectors().similar_by_vector(query_vector, 50)
@@ -41,20 +42,25 @@ def do_cvd_search(search_run: SearchRun):
     wordform_queries = [
         cvd_key_to_wordform_query(similarity) for similarity, weight in closest
     ]
+    print(wordform_queries)
     similarities = [similarity for cvd_key, similarity in closest]
-
+    print(similarities)
     # Get all possible wordforms in one big query. We will select more than we
     # need, then filter it down later, but this will have to do until we get
     # better homonym handling.
+    print(Wordform.objects.count())
     wordform_results = Wordform.objects.filter(
         text__in=set(wf["text"] for wf in wordform_queries)
     )
+    print(wordform_results)
 
     # Now match back up
     wordforms_by_text = {
         text: list(wordforms)
         for text, wordforms in itertools.groupby(wordform_results, key=lambda x: x.text)
     }
+
+    print(wordforms_by_text)
 
     for similarity, wordform_query in zip(similarities, wordform_queries):
         # gensim uses the terminology, similarity = 1 - distance. Its
@@ -64,6 +70,7 @@ def do_cvd_search(search_run: SearchRun):
         distance = 1 - similarity
 
         wordforms_for_query = wordforms_by_text.get(wordform_query["text"], None)
+        print(wordforms_for_query)
         if wordforms_for_query is None:
             logger.warning(
                 f"Wordform {wordform_query['text']} not found in CVD; mismatch between definition vector model file and definitions in database?"

src/views.py

@@ -3,7 +3,7 @@
 import json
 import logging
 from pathlib import Path
-
+import time
 from typing import Dict, Literal
 from nltk.corpus import wordnet as wn
 
@@ -187,6 +187,7 @@ def search_api(request):
     :param request:
     :return:
     """
+    start = time.time()
     query_string = request.GET.get("name")
     rw_index = request.GET.get("rw_index")
     rw_domain = request.GET.get("rw_domain")
@@ -227,10 +228,6 @@ def search_api(request):
             search_run.verbose_messages, indent=2, ensure_ascii=False
         )
 
-    context["search_results"] = fetch_single_recording(
-        context["search_results"], request
-    )
-
     for result in context["search_results"]:
         result["wordform_text"] = wordform_orth_text(result["wordform_text"])
         result["lemma_wordform"]["wordform_text"] = wordform_orth_text(
@@ -246,7 +243,8 @@ def search_api(request):
             result["relabelled_fst_analysis"] = relabelFSTAnalysis(
                 result["relabelled_fst_analysis"]
             )
-
+    end = time.time()
+    print(end - start)
     return Response(context)
 
 
@@ -300,29 +298,6 @@ def wordnet_api(request, classification):
     return Response(context)
 
 
-def fetch_single_recording(results, request):
-    query_terms = []
-    for result in results:
-        query_terms.append(result["wordform_text"])
-
-    speech_db_eq = settings.SPEECH_DB_EQ
-    matched_recordings = {}
-
-    for search_terms in divide_chunks(query_terms, 30):
-        for source in speech_db_eq:
-            url = f"https://speech-db.altlab.app/{source}/api/bulk_search"
-            matched_recordings.update(get_recordings_from_url(search_terms, url))
-
-    for result in results:
-        if result["wordform_text"] in matched_recordings:
-            result["recording"] = matched_recordings[result["wordform_text"]][
-                "recording_url"
-            ]
-        else:
-            result["recording"] = ""
-
-    return results
-
 
 def relabelInflectionalCategory(ic):
     with open(Path(settings.RESOURCES_DIR / "altlabel.tsv")) as f: