Kuberwastaken commited on
Commit
c8c52e3
·
1 Parent(s): 07f4e87

Improved detection

Browse files
Files changed (1) hide show
  1. model/analyzer.py +12 -8
model/analyzer.py CHANGED
@@ -90,7 +90,7 @@ class ContentAnalyzer:
90
  prompt = f"""
91
  Check this text for any indication of {mapped_name} ({description}).
92
  Be sensitive to subtle references or implications, make sure the text is not metaphorical.
93
- Respond concisely with: YES, NO, or MAYBE.
94
  Text: {chunk}
95
  Answer:
96
  """
@@ -112,17 +112,21 @@ class ContentAnalyzer:
112
  )
113
 
114
  response_text = self.tokenizer.decode(outputs[0], skip_special_tokens=True).strip().upper()
115
- first_word = response_text.split("\n")[-1].split()[0] if response_text else "NO"
116
- print(f"Model response for {mapped_name}: {first_word}")
117
 
118
- if first_word == "YES":
119
- print(f"Detected {mapped_name} in this chunk!")
 
120
  chunk_triggers[mapped_name] = chunk_triggers.get(mapped_name, 0) + 1
121
- elif first_word == "MAYBE":
122
- print(f"Possible {mapped_name} detected, marking for further review.")
123
  chunk_triggers[mapped_name] = chunk_triggers.get(mapped_name, 0) + 0.5
124
  else:
125
- print(f"No {mapped_name} detected in this chunk.")
 
 
 
 
 
126
 
127
  if progress:
128
  current_progress += progress_step
 
90
  prompt = f"""
91
  Check this text for any indication of {mapped_name} ({description}).
92
  Be sensitive to subtle references or implications, make sure the text is not metaphorical.
93
+ Respond concisely and ONLY with: YES, NO, or MAYBE.
94
  Text: {chunk}
95
  Answer:
96
  """
 
112
  )
113
 
114
  response_text = self.tokenizer.decode(outputs[0], skip_special_tokens=True).strip().upper()
 
 
115
 
116
+ # Scan the entire response for the relevant words
117
+ if "YES" in response_text:
118
+ print(f"{' ' * 16}Detected {mapped_name} in this chunk!")
119
  chunk_triggers[mapped_name] = chunk_triggers.get(mapped_name, 0) + 1
120
+ elif "MAYBE" in response_text:
121
+ print(f"{' ' * 16}Possible {mapped_name} detected, marking for further review.")
122
  chunk_triggers[mapped_name] = chunk_triggers.get(mapped_name, 0) + 0.5
123
  else:
124
+ print(f"{' ' * 16}No {mapped_name} detected in this chunk.")
125
+
126
+ if progress:
127
+ current_progress += progress_step
128
+ progress(min(current_progress, 0.9), f"{' ' * 16}Analyzing {mapped_name}...")
129
+
130
 
131
  if progress:
132
  current_progress += progress_step