Create app.py

app.py

@@ -0,0 +1,68 @@
+import gradio as gr
+from sentence_transformers import SentenceTransformer
+from sklearn.cluster import KMeans
+import pandas as pd
+import plotly.express as px
+import umap
+
+model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
+
+def analyze_problems(text):
+
+    problems = [p.strip() for p in text.split("\n") if p.strip()]
+
+    embeddings = model.encode(problems)
+
+    # clustering
+    k = min(5, len(problems))
+    kmeans = KMeans(n_clusters=k, random_state=0).fit(embeddings)
+
+    clusters = kmeans.labels_
+
+    # dimensionality reduction for visualization
+    reducer = umap.UMAP()
+    coords = reducer.fit_transform(embeddings)
+
+    df = pd.DataFrame({
+        "problem": problems,
+        "cluster": clusters,
+        "x": coords[:,0],
+        "y": coords[:,1]
+    })
+
+    fig = px.scatter(
+        df,
+        x="x",
+        y="y",
+        color=df["cluster"].astype(str),
+        text="problem",
+        title="Problem Market Map"
+    )
+
+    cluster_summary = df.groupby("cluster")["problem"].apply(list).to_dict()
+
+    summary_text = ""
+
+    for c, items in cluster_summary.items():
+        summary_text += f"\nCluster {c}\n"
+        for i in items:
+            summary_text += f"- {i}\n"
+
+    return summary_text, fig
+
+
+with open("sample_problems.txt") as f:
+    default_text = f.read()
+
+demo = gr.Interface(
+    fn=analyze_problems,
+    inputs=gr.Textbox(value=default_text, lines=15, label="Problem Signals"),
+    outputs=[
+        gr.Textbox(label="Problem Clusters"),
+        gr.Plot(label="Problem Market Map")
+    ],
+    title="Problem Discovery Engine Demo",
+    description="Detect clusters of real-world problems and visualize a Problem Market Map."
+)
+
+demo.launch()