all the project copy

api/phrase_similarity.py

@@ -0,0 +1,64 @@
+from pathlib import Path
+import sys
+import traceback
+
+
+ROOT_DIR = Path(__file__).resolve().parents[1]
+CONN_DIR = ROOT_DIR / "conn"
+if str(CONN_DIR) not in sys.path:
+    sys.path.insert(0, str(CONN_DIR))
+
+import mysql5
+
+
+def _legacy_start_guessing_manual(ticket):
+    try:
+        from sklearn.feature_extraction.text import TfidfVectorizer
+        from sklearn.metrics.pairwise import cosine_similarity
+    except Exception:
+        traceback.print_exc()
+        return []
+
+    competence_origin, product_origin, problem_origin, area_origin, subarea_origin = mysql5.getCompetenceOrProdotto(ticket)
+    w_list = mysql5.getProblemaBYCompetenzaOrProgramma_manul(competence_origin, product_origin, area_origin, subarea_origin)
+
+    if len(w_list) == 0:
+        return []
+
+    phrases = [item[0] for item in w_list]
+    vectorizer = TfidfVectorizer(analyzer="char_wb", ngram_range=(3, 5), lowercase=True)
+    tfidf_matrix = vectorizer.fit_transform([problem_origin] + phrases)
+    similarity_matrix = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:])
+
+    # Keep the real top five. The old code used [-6:-1], which skipped the best result.
+    best_matches_indices = similarity_matrix.argsort()[0][-5:][::-1]
+
+    final_list = []
+    for i, index in enumerate(best_matches_indices, start=1):
+        complete_ticket = mysql5.getTicketByProblema(str(phrases[index]))
+        solution = str(complete_ticket[1])
+        if len(solution) > 400:
+            solution = solution[:400] + "..."
+
+        final_list.append(
+            f"\n\n----- {i}. Metodo legacy -----\n"
+            f"[ /Ticket_dettaglio_{complete_ticket[0]} ]\n"
+            f"https://tsnew.sanmarcoweb.com/it/ticket/index/index/operation/view/id/{complete_ticket[0]}\n"
+            f"SOLUZIONE: {solution}\n"
+        )
+    return final_list
+
+
+def start_guessing_manual(ticket, user_id=None, filters=None, limit=None, offset=0):
+    try:
+        import ai_retriever
+
+        kwargs = {"user_id": user_id, "filters": filters, "offset": offset}
+        if limit is not None:
+            kwargs["limit"] = limit
+        return ai_retriever.start_guessing_manual(ticket, **kwargs)
+    except Exception:
+        traceback.print_exc()
+        if filters:
+            return []
+        return _legacy_start_guessing_manual(ticket)