momenaca commited on
Commit
dd621e1
·
1 Parent(s): 185af4d

update politique specific database

Browse files
Files changed (2) hide show
  1. app.py +21 -1
  2. spinoza_project/source/frontend/utils.py +16 -0
app.py CHANGED
@@ -15,6 +15,7 @@ from spinoza_project.source.frontend.utils import (
15
  make_html_source,
16
  make_html_presse_source,
17
  make_html_afp_source,
 
18
  parse_output_llm_with_sources,
19
  init_env,
20
  )
@@ -53,9 +54,10 @@ bdd_afp = get_vectorstore_api("afp")
53
  qdrants = {
54
  tab: pickle_to_document_store(
55
  hf_hub_download(
56
- repo_id="SpinozaProject/spinoza-database",
57
  filename=f"database_{tab}.pickle",
58
  repo_type="dataset",
 
59
  )
60
  )
61
  for tab in config["prompt_naming"]
@@ -243,6 +245,24 @@ def get_sources(
243
  ]
244
  )
245
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
246
  else:
247
  sources = qdrants[
248
  config["source_mapping"][tab]
 
15
  make_html_source,
16
  make_html_presse_source,
17
  make_html_afp_source,
18
+ make_html_politique_source,
19
  parse_output_llm_with_sources,
20
  init_env,
21
  )
 
54
  qdrants = {
55
  tab: pickle_to_document_store(
56
  hf_hub_download(
57
+ repo_id="TestSpinoza/spinoza-database",
58
  filename=f"database_{tab}.pickle",
59
  repo_type="dataset",
60
+ force_download=True,
61
  )
62
  )
63
  for tab in config["prompt_naming"]
 
245
  ]
246
  )
247
 
248
+ elif tab == "Documents Stratégiques":
249
+ sources = qdrants[
250
+ config["source_mapping"][tab]
251
+ ].similarity_search_with_relevance_scores(
252
+ config["query_preprompt"]
253
+ + question.replace("<p>", "").replace("</p>\n", ""),
254
+ k=k,
255
+ )
256
+ sources = [
257
+ (doc, score) for doc, score in sources if score >= min_similarity
258
+ ]
259
+ formated.extend(
260
+ [
261
+ make_html_politique_source(source[0], j, source[1], config)
262
+ for j, source in zip(range(k * i + 1, k * (i + 1) + 1), sources)
263
+ ]
264
+ )
265
+
266
  else:
267
  sources = qdrants[
268
  config["source_mapping"][tab]
spinoza_project/source/frontend/utils.py CHANGED
@@ -107,6 +107,22 @@ def make_html_afp_source(source, i, score):
107
  """
108
 
109
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
110
  def make_html_source(source, i, score, config):
111
  meta = source.metadata
112
  return f"""
 
107
  """
108
 
109
 
110
+ def make_html_politique_source(source, i, score, config):
111
+ meta = source.metadata
112
+ return f"""
113
+ <div class="card" id="doc{i}">
114
+ <div class="card-content">
115
+ <h2>Doc {i} - {meta['file_name'].replace(".pdf", "").replace("_", " ").replace("-", " ")} - Page {meta['content_page_number'] + 1}</h2>
116
+ <p>{source.page_content.replace(config["passage_preprompt"], "")}</p>
117
+ </div>
118
+ <div class="card-footer">
119
+ <span>Politique - {meta['file_source_type']}</span>
120
+ <span>Relevance Score : {round(100*score,1)}%</span>
121
+ </div>
122
+ </div>
123
+ """
124
+
125
+
126
  def make_html_source(source, i, score, config):
127
  meta = source.metadata
128
  return f"""