Update app.py

app.py

@@ -3,7 +3,7 @@ import pandas as pd
 from transformers import pipeline
 
 # ------------------------------------------------
-# Load Small Language Model (light polishing only)
+# Load SLM (polishing only)
 # ------------------------------------------------
 generator = pipeline(
     "text2text-generation",
@@ -11,7 +11,7 @@ generator = pipeline(
 )
 
 # ------------------------------------------------
-# Generic language helpers (NO hard coding)
+# Generic helpers (pure statistics → language)
 # ------------------------------------------------
 def trend_word(diff):
     return "increased" if diff > 0 else "decreased"
@@ -27,19 +27,41 @@ def magnitude_word(abs_diff):
 def unit_word(kpi):
     return "percentage points" if "%" in kpi else ""
 
+def summarize_secondary(diffs):
+    avg_change = diffs.abs().mean()
+    inc_ratio = (diffs > 0).mean()
+
+    # Magnitude descriptor
+    if avg_change < 0.05:
+        magnitude = "minor variation"
+    elif avg_change < 0.2:
+        magnitude = "moderate movement"
+    else:
+        magnitude = "notable movement"
+
+    # Direction descriptor
+    if inc_ratio > 0.7:
+        direction = "mostly increased"
+    elif inc_ratio < 0.3:
+        direction = "mostly decreased"
+    else:
+        direction = "mixed movement"
+
+    return f"Remaining indicators showed {magnitude} with {direction}"
+
 # ------------------------------------------------
 # Core logic
 # ------------------------------------------------
 def analyze_kpi(csv_file, top_n):
     df = pd.read_csv(csv_file.name)
 
-    # Identify date columns
     date_cols = df.columns[1:]
     prev_date = date_cols[-2]
     curr_date = date_cols[-1]
 
-    # Compute absolute change
-    df["Change"] = (df[curr_date] - df[prev_date]).abs()
+    df["Diff"] = df[curr_date] - df[prev_date]
+    df["Change"] = df["Diff"].abs()
+
     df_sorted = df.sort_values("Change", ascending=False)
     top_kpis = df_sorted.head(top_n)
 
@@ -47,28 +69,25 @@ def analyze_kpi(csv_file, top_n):
     # Primary KPI sentence
     # -------------------------------
     primary = top_kpis.iloc[0]
-    diff = primary[curr_date] - primary[prev_date]
-    abs_diff = abs(diff)
+    abs_diff = abs(primary["Diff"])
 
     primary_sentence = (
         f"{primary['Kpi']} "
-        f"{trend_word(diff)} "
+        f"{trend_word(primary['Diff'])} "
         f"{magnitude_word(abs_diff)} "
         f"by {abs_diff:.2f} {unit_word(primary['Kpi'])}".strip()
     )
 
     # -------------------------------
-    # Secondary KPIs sentence
+    # Secondary KPIs sentence (DATA-DRIVEN)
     # -------------------------------
-    secondary_sentence = (
-        "Other key indicators showed only minor variation, "
-        "suggesting overall stability across the remaining metrics."
-    )
+    secondary_diffs = top_kpis.iloc[1:]["Diff"]
+    secondary_sentence = summarize_secondary(secondary_diffs)
 
-    # Input to SLM (already well-formed)
-    model_input = primary_sentence + ". " + secondary_sentence
+    # Combined input
+    model_input = primary_sentence + ". " + secondary_sentence + "."
 
-    # Light polishing only
+    # Light smoothing only
     summary = generator(
         model_input,
         max_new_tokens=45,
@@ -78,43 +97,27 @@ def analyze_kpi(csv_file, top_n):
     return top_kpis[["Kpi", "Change"]], summary
 
 # ------------------------------------------------
-# Gradio UI (Hugging Face Space)
+# Gradio UI (HF Space)
 # ------------------------------------------------
 with gr.Blocks(title="KPI Change Newsletter") as demo:
     gr.Markdown("## 📰 KPI Change Newsletter Generator")
     gr.Markdown(
-        "Upload a KPI CSV file to automatically rank changes "
-        "and generate a short, readable operational summary."
-    )
-
-    csv_input = gr.File(
-        label="Upload CSV",
-        file_types=[".csv"]
+        "Upload a KPI CSV file to rank changes and generate "
+        "a short, data-driven summary."
     )
 
-    top_n_input = gr.Slider(
-        minimum=3,
-        maximum=5,
-        value=3,
-        step=1,
-        label="Number of KPIs to Rank"
-    )
+    csv_input = gr.File(label="Upload CSV", file_types=[".csv"])
+    top_n_input = gr.Slider(3, 5, value=3, step=1, label="Top KPIs")
 
     generate_btn = gr.Button("Generate Summary")
 
-    table_output = gr.Dataframe(
-        label="Top KPI Changes"
-    )
-
-    summary_output = gr.Textbox(
-        label="Newsletter Summary",
-        lines=2
-    )
+    table_output = gr.Dataframe(label="Top KPI Changes")
+    summary_output = gr.Textbox(label="Newsletter Summary", lines=2)
 
     generate_btn.click(
-        fn=analyze_kpi,
+        analyze_kpi,
         inputs=[csv_input, top_n_input],
         outputs=[table_output, summary_output]
     )
 
-demo.launch()
+demo.launch()