Update app.py
Browse files
app.py
CHANGED
@@ -301,12 +301,10 @@ for key in ['pdf_processed', 'markdown_texts', 'df']:
|
|
301 |
# ---------------------------------------------------------------------------------------
|
302 |
# API Configuration
|
303 |
# ---------------------------------------------------------------------------------------
|
304 |
-
# Retrieve Hugging Face API key from environment variables
|
305 |
hf_api_key = os.getenv('HF_API_KEY')
|
306 |
if not hf_api_key:
|
307 |
raise ValueError("HF_API_KEY not set in environment variables")
|
308 |
|
309 |
-
# Create the Hugging Face inference client
|
310 |
client = InferenceClient(api_key=hf_api_key)
|
311 |
|
312 |
# ---------------------------------------------------------------------------------------
|
@@ -321,8 +319,8 @@ class SurveyAnalysis:
|
|
321 |
Instructions:
|
322 |
- Extract exact quotes per topic.
|
323 |
- Ignore irrelevant topics.
|
|
|
324 |
|
325 |
-
Format:
|
326 |
[Topic]
|
327 |
- "Exact quote"
|
328 |
|
@@ -331,32 +329,31 @@ Meeting Notes:
|
|
331 |
"""
|
332 |
|
333 |
def prompt_response_from_hf_llm(self, llm_input):
|
334 |
-
# Define a system prompt to guide the model's responses
|
335 |
system_prompt = """
|
336 |
-
|
337 |
-
|
338 |
-
|
339 |
-
|
|
|
|
|
|
|
|
|
|
|
340 |
"""
|
341 |
-
|
342 |
-
# Generate the refined prompt using Hugging Face API
|
343 |
response = client.chat.completions.create(
|
344 |
model="meta-llama/Llama-3.1-70B-Instruct",
|
345 |
messages=[
|
346 |
-
{"role": "system", "content": system_prompt},
|
347 |
{"role": "user", "content": llm_input}
|
348 |
],
|
349 |
-
stream=True,
|
350 |
temperature=0.5,
|
351 |
max_tokens=1024,
|
352 |
top_p=0.7
|
353 |
)
|
354 |
-
|
355 |
-
# Combine messages if response is streamed
|
356 |
-
response_content = ""
|
357 |
-
for message in response:
|
358 |
-
response_content += message.choices[0].delta.content
|
359 |
|
|
|
|
|
360 |
return response_content.strip()
|
361 |
|
362 |
def extract_text(self, response):
|
@@ -367,7 +364,6 @@ Meeting Notes:
|
|
367 |
for _, row in df.iterrows():
|
368 |
llm_input = self.prepare_llm_input(row['Document_Text'], topics)
|
369 |
response = self.prompt_response_from_hf_llm(llm_input)
|
370 |
-
print("AI Response:", response) # Debugging: print the AI response
|
371 |
notes = self.extract_text(response)
|
372 |
results.append({'Document_Text': row['Document_Text'], 'Topic_Summary': notes})
|
373 |
return pd.concat([df.reset_index(drop=True), pd.DataFrame(results)['Topic_Summary']], axis=1)
|
@@ -408,17 +404,24 @@ def extract_markdown_from_image(image):
|
|
408 |
doc.load_from_doctags(doctags_doc)
|
409 |
return doc.export_to_markdown()
|
410 |
|
|
|
411 |
def extract_excerpts(processed_df):
|
412 |
rows = []
|
413 |
for _, r in processed_df.iterrows():
|
414 |
-
|
415 |
-
|
|
|
416 |
if topic_match:
|
417 |
-
topic = topic_match.group(1)
|
418 |
-
excerpts = re.findall(r'- "([^"]+)"', sec)
|
419 |
for excerpt in excerpts:
|
420 |
-
rows.append({
|
421 |
-
|
|
|
|
|
|
|
|
|
|
|
422 |
return pd.DataFrame(rows)
|
423 |
|
424 |
# ---------------------------------------------------------------------------------------
|
|
|
301 |
# ---------------------------------------------------------------------------------------
|
302 |
# API Configuration
|
303 |
# ---------------------------------------------------------------------------------------
|
|
|
304 |
hf_api_key = os.getenv('HF_API_KEY')
|
305 |
if not hf_api_key:
|
306 |
raise ValueError("HF_API_KEY not set in environment variables")
|
307 |
|
|
|
308 |
client = InferenceClient(api_key=hf_api_key)
|
309 |
|
310 |
# ---------------------------------------------------------------------------------------
|
|
|
319 |
Instructions:
|
320 |
- Extract exact quotes per topic.
|
321 |
- Ignore irrelevant topics.
|
322 |
+
- Strictly follow this format:
|
323 |
|
|
|
324 |
[Topic]
|
325 |
- "Exact quote"
|
326 |
|
|
|
329 |
"""
|
330 |
|
331 |
def prompt_response_from_hf_llm(self, llm_input):
|
|
|
332 |
system_prompt = """
|
333 |
+
You are an expert assistant tasked with extracting exact quotes from provided meeting notes based on given topics.
|
334 |
+
|
335 |
+
Instructions:
|
336 |
+
- Only extract exact quotes relevant to provided topics.
|
337 |
+
- Ignore irrelevant content.
|
338 |
+
- Strictly follow this format:
|
339 |
+
|
340 |
+
[Topic]
|
341 |
+
- "Exact quote"
|
342 |
"""
|
343 |
+
|
|
|
344 |
response = client.chat.completions.create(
|
345 |
model="meta-llama/Llama-3.1-70B-Instruct",
|
346 |
messages=[
|
347 |
+
{"role": "system", "content": system_prompt},
|
348 |
{"role": "user", "content": llm_input}
|
349 |
],
|
|
|
350 |
temperature=0.5,
|
351 |
max_tokens=1024,
|
352 |
top_p=0.7
|
353 |
)
|
|
|
|
|
|
|
|
|
|
|
354 |
|
355 |
+
response_content = response.choices[0].message.content
|
356 |
+
print("Full AI Response:", response_content) # Debugging
|
357 |
return response_content.strip()
|
358 |
|
359 |
def extract_text(self, response):
|
|
|
364 |
for _, row in df.iterrows():
|
365 |
llm_input = self.prepare_llm_input(row['Document_Text'], topics)
|
366 |
response = self.prompt_response_from_hf_llm(llm_input)
|
|
|
367 |
notes = self.extract_text(response)
|
368 |
results.append({'Document_Text': row['Document_Text'], 'Topic_Summary': notes})
|
369 |
return pd.concat([df.reset_index(drop=True), pd.DataFrame(results)['Topic_Summary']], axis=1)
|
|
|
404 |
doc.load_from_doctags(doctags_doc)
|
405 |
return doc.export_to_markdown()
|
406 |
|
407 |
+
# Revised extract_excerpts function with improved robustness
|
408 |
def extract_excerpts(processed_df):
|
409 |
rows = []
|
410 |
for _, r in processed_df.iterrows():
|
411 |
+
sections = re.split(r'\n(?=(?:\*\*|\[)?[A-Za-z/ ]+(?:\*\*|\])?\n- )', r['Topic_Summary'])
|
412 |
+
for sec in sections:
|
413 |
+
topic_match = re.match(r'(?:\*\*|\[)?([A-Za-z/ ]+)(?:\*\*|\])?', sec.strip())
|
414 |
if topic_match:
|
415 |
+
topic = topic_match.group(1).strip()
|
416 |
+
excerpts = re.findall(r'- "?([^"\n]+)"?', sec)
|
417 |
for excerpt in excerpts:
|
418 |
+
rows.append({
|
419 |
+
'Document_Text': r['Document_Text'],
|
420 |
+
'Topic_Summary': r['Topic_Summary'],
|
421 |
+
'Excerpt': excerpt.strip(),
|
422 |
+
'Topic': topic
|
423 |
+
})
|
424 |
+
print("Extracted Rows:", rows) # Debugging
|
425 |
return pd.DataFrame(rows)
|
426 |
|
427 |
# ---------------------------------------------------------------------------------------
|