cleaned control file

src/control.py

@@ -1,49 +0,0 @@
-import chromadb
-
-import src.tools.retriever as rtrvr
-import src.tools.llm as llm
-from src.domain.doc import Doc
-
-chroma_client = chromadb.Client()
-
-plan_language = 'en'
-content_language = 'en'
-path_plan = '../data/business_trips_plan_until_9_en.docx'
-path_content = '../data/business_trips_content_until_9_en.docx'
-collection_name = "until_9"
-
-doc_plan = Doc(path_plan)
-doc_content = Doc(path_content)
-collection_ = rtrvr.init_collections(chroma_client, doc_plan, doc_content, collection_name)
-
-
-def get_response(query):
-    if plan_language == 'en':
-        query = llm.translate(query)
-    sources = rtrvr.similarity_search(collection=collection_, query=query)
-    sources = select_best_sources(sources)
-    sources_contents = [s['content'] for s in sources]
-    context = '\n'.join(sources_contents)
-    answer = llm.generate_paragraph(query=query, context=context, language=content_language)
-    if content_language == 'en':
-        answer = llm.translate(text=answer, language='fr')
-    return answer.lstrip(), sources
-
-
-def select_best_sources(sources: [], delta_1_2=0.1, delta_1_n=0.25, absolute=1.1) -> []:
-    best_sources = []
-    for idx, s in enumerate(sources):
-        if idx == 0 \
-           or (s['distance_f'] - sources[idx - 1]['distance_f'] < delta_1_2
-                and s['distance_f'] - sources[0]['distance_f'] < delta_1_n) \
-           or s['distance_f'] < absolute:
-            best_sources.append(s)
-    return best_sources
-
-
-q1 = "Comment sont remboursés mes frais kilométriques sur mes déplacements avec mon véhicule personnel?"
-q2 = "Quels sont les moyens de transport recommandés par la société?"
-q3 = "est-ce que mes billets de cinéma peuvent être remboursés?"
-
-a2 = get_response(q3)
-print(a2)