Update app.py

app.py

@@ -7,14 +7,6 @@ import easyocr
 import numpy as np
 import pandas as pd
 
-mpqa = {}
-with open("subj_lexicon.tff") as f:
-    for line in f:
-        fields = dict(tok.split("=") for tok in line.strip().split())
-        word = fields["word1"]
-        strength = fields["type"]            # “strongsubj” vs “weaksubj”
-        polarity = fields["priorpolarity"]   # “positive” or “negative”
-        mpqa[word] = (strength, polarity)
 
 # ——— Load and preprocess NRC EmoLex ——————————————————————————————————
 # Make sure this filename matches exactly what you’ve uploaded
@@ -47,7 +39,47 @@ def score_emolex(text_lower):
             for emo, flag in EMOLEX[tok].items():
                 counts[emo] += flag
     return counts
+import re
+
+# ——— Load MPQA Subjectivity Lexicon —————————————————————————————————————————————
+MPQA_PATH = "subjclueslen1-HLTEMNLP05.tff"
 
+# mpqa_lex[word] = list of feature‐dicts for that word
+mpqa_lex = {}
+with open(MPQA_PATH, encoding="utf-8") as f:
+    for line in f:
+        line = line.strip()
+        if not line or line.startswith("#"):
+            continue
+        # each line looks like: type=strongsubj len=1 word1=abandon pos1=verb stemmed1=y priorpolarity=negative
+        fields = dict(item.split("=",1) for item in line.split())
+        w = fields.pop("word1").lower()
+        mpqa_lex.setdefault(w, []).append(fields)
+
+# e.g. mpqa_lex["abandon"] == [ {'type':'strongsubj','len':'1','pos1':'verb','stemmed1':'y','priorpolarity':'negative'} ]
+
+# ——— In your get_emotional_tone_tag, just after you split words… ——————————————————————
+words = text_lower.split()
+
+# count MPQA hits
+mpqa_counts = {
+    "strongsubj":  0,
+    "weaksubj":    0,
+    "positive":    0,
+    "negative":    0,
+}
+for w in words:
+    for entry in mpqa_lex.get(w, []):
+        mpqa_counts[ entry["type"] ]      += 1
+        mpqa_counts[ entry["priorpolarity"] ] += 1
+
+# now you can reference mpqa_counts["negative"], etc.
+# for example, tweak your “Emotional Threat” rule to require at least one strong negative subj:
+if (anger + disgust) > 0.5 \
+   and (lex_counts["anger"] > 0 or lex_counts["disgust"] > 0) \
+   and mpqa_counts["strongsubj"] > 0 \
+   and any(p in patterns for p in ["control","threat","insults","dismissiveness"]):
+    return "emotional threat"
 # ——— 1) Emotion Pipeline ————————————————————————————————————————————————
 emotion_pipeline = hf_pipeline(
     "text-classification",