Update app.py

app.py

@@ -40,7 +40,18 @@ with open(MPQA_PATH, encoding="utf-8") as f:
         line = line.strip()
         if not line or line.startswith("#"):
             continue
-        fields = dict(item.split("=",1) for item in line.split())
+
+        # build fields dict but skip any token without '='
+        fields = {}
+        for item in line.split():
+            if "=" not in item:
+                continue
+            key, val = item.split("=", 1)
+            fields[key] = val
+
+        # must have word1
+        if "word1" not in fields:
+            continue
         w = fields.pop("word1").lower()
         mpqa_lex.setdefault(w, []).append(fields)
 
@@ -88,11 +99,6 @@ ocr_reader = easyocr.Reader(["en"], gpu=False)
 
 # ——— 4) Emotional-Tone Tagging —————————————————————————————————————————————
 def get_emotional_tone_tag(emotion_profile, patterns, text_lower):
-    """
-    Assigns one of 18 nuanced tone categories based on
-    model scores, NRC-EmoLex counts, MPQA counts, detected patterns, and text.
-    """
-    # unpack transformer scores
     sadness  = emotion_profile.get("sadness",  0)
     joy      = emotion_profile.get("joy",      0)
     neutral  = emotion_profile.get("neutral",  0)
@@ -234,15 +240,15 @@ def analyze_message(text):
     text_lower = text.lower()
     emotion_profile = get_emotion_profile(text)
 
-    # blend in NRC-EmoLex scores
+    # blend in NRC-EmoLex
     lex_counts = score_emolex(text_lower)
     max_lex    = max(lex_counts.values()) or 1.0
-    lex_scores = {emo: cnt/ max_lex for emo, cnt in lex_counts.items()}
+    lex_scores = {emo: cnt/max_lex for emo, cnt in lex_counts.items()}
     for emo in emotion_profile:
         emotion_profile[emo] = max(emotion_profile[emo], lex_scores.get(emo,0))
 
     # abuse-patterns
-    toks = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
+    toks   = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
     with torch.no_grad():
         logits = model(**toks).logits.squeeze(0)
     scores = torch.sigmoid(logits).cpu().numpy()
@@ -261,7 +267,7 @@ def analyze_message(text):
 def analyze_composite(uploaded_file, *texts):
     outputs = []
 
-    # file OCR / text handling
+    # file handling / OCR
     if uploaded_file is not None:
         try:
             raw = uploaded_file.read()
@@ -271,8 +277,8 @@ def analyze_composite(uploaded_file, *texts):
 
         name = uploaded_file.name.lower() if hasattr(uploaded_file,"name") else uploaded_file.lower()
         if name.endswith((".png",".jpg",".jpeg",".bmp",".gif",".tiff")):
-            img = Image.open(io.BytesIO(raw))
-            arr = np.array(img.convert("RGB"))
+            img     = Image.open(io.BytesIO(raw))
+            arr     = np.array(img.convert("RGB"))
             content = "\n".join(ocr_reader.readtext(arr, detail=0))
         else:
             try:
@@ -288,7 +294,7 @@ def analyze_composite(uploaded_file, *texts):
             f"Emotional Tone  : {r['tone_tag']}\n"
         )
 
-    # inline text messages
+    # free-text messages
     for idx, txt in enumerate(texts, start=1):
         if not txt:
             continue
@@ -306,6 +312,7 @@ def analyze_composite(uploaded_file, *texts):
 
 # ——— 7) Gradio interface ———————————————————————————————————————————————
 message_inputs = [gr.Textbox(label="Message")]
+
 iface = gr.Interface(
     fn=analyze_composite,
     inputs=[gr.File(file_types=[".txt",".png",".jpg",".jpeg"], label="Upload text or image")] + message_inputs,