add visit counter

bde_prediction/bde_flask/__init__.py

@@ -8,6 +8,7 @@ from alfabet.prediction import predict_bdes, check_input
 from flask import Flask, render_template, request, flash, jsonify, send_from_directory
 from werkzeug.middleware.dispatcher import DispatcherMiddleware
 from wtforms import Form, StringField, validators
+from visit_counter import log_visit
 
 # App config.
 
@@ -57,6 +58,7 @@ def send_static(path):
 
 @app.route("/", methods=['GET', 'POST'])
 def index():
+    log_visit()
     form = ReusableForm(request.form)
     return render_template('index.html', form=form)
 

bde_prediction/visit_counter.py

@@ -0,0 +1,139 @@
+"""
+visit_counter.py
+----------------
+Drop this file into your Space. Call `log_visit()` once per app load.
+
+Setup:
+1. Create a private HF Dataset repo for analytics:
+      huggingface-cli repo create analytics --type dataset
+
+2. Add your HF token as a Space secret:
+      HF_TOKEN = <your token with write access>
+
+3. Call log_visit() at app startup (see bottom of file for example).
+"""
+
+import os
+import csv
+import io
+from datetime import datetime, timezone
+
+try:
+    from huggingface_hub import HfApi, hf_hub_download
+    from huggingface_hub.utils import EntryNotFoundError
+except ImportError:
+    import subprocess, sys
+    subprocess.check_call([sys.executable, "-m", "pip", "install", "huggingface_hub", "-q"])
+    from huggingface_hub import HfApi, hf_hub_download
+    from huggingface_hub.utils import EntryNotFoundError
+
+
+# ── Config ────────────────────────────────────────────────────────────────────
+ANALYTICS_REPO = "patonlab/analytics"   # your private dataset repo
+ANALYTICS_FILE = "visits.csv"           # filename inside that repo
+SPACE_NAME     = "patonlab/alfabet_bde" # label stored in each row
+TOKEN          = os.environ.get("HF_TOKEN")  # set as a Space secret
+# ─────────────────────────────────────────────────────────────────────────────
+
+
+def log_visit(space: str = SPACE_NAME):
+    """Append one row to visits.csv in the analytics dataset repo."""
+    if not TOKEN:
+        print("[visit_counter] HF_TOKEN not set — skipping visit log.")
+        return
+
+    api = HfApi(token=TOKEN)
+    now = datetime.now(timezone.utc).isoformat()
+
+    # 1. Try to download existing CSV
+    existing_rows = []
+    try:
+        path = hf_hub_download(
+            repo_id=ANALYTICS_REPO,
+            filename=ANALYTICS_FILE,
+            repo_type="dataset",
+            token=TOKEN,
+        )
+        with open(path, newline="") as f:
+            existing_rows = list(csv.DictReader(f))
+    except EntryNotFoundError:
+        pass  # first visit ever — file doesn't exist yet
+
+    # 2. Append new row
+    existing_rows.append({"space": space, "timestamp": now})
+
+    # 3. Write updated CSV to buffer
+    buf = io.StringIO()
+    writer = csv.DictWriter(buf, fieldnames=["space", "timestamp"])
+    writer.writeheader()
+    writer.writerows(existing_rows)
+
+    # 4. Push back to dataset repo
+    api.upload_file(
+        path_or_fileobj=buf.getvalue().encode(),
+        path_in_repo=ANALYTICS_FILE,
+        repo_id=ANALYTICS_REPO,
+        repo_type="dataset",
+        commit_message=f"visit: {space} @ {now}",
+    )
+    print(f"[visit_counter] Logged visit for {space} at {now}")
+
+
+def get_visit_counts() -> dict:
+    """Return a dict of {space_name: visit_count} from the analytics repo."""
+    if not TOKEN:
+        return {}
+    try:
+        path = hf_hub_download(
+            repo_id=ANALYTICS_REPO,
+            filename=ANALYTICS_FILE,
+            repo_type="dataset",
+            token=TOKEN,
+        )
+        counts = {}
+        with open(path, newline="") as f:
+            for row in csv.DictReader(f):
+                counts[row["space"]] = counts.get(row["space"], 0) + 1
+        return counts
+    except EntryNotFoundError:
+        return {}
+
+
+def print_summary():
+    """Print a quick analytics summary to stdout."""
+    counts = get_visit_counts()
+    if not counts:
+        print("No visit data yet.")
+        return
+    total = sum(counts.values())
+    print(f"\n{'='*40}")
+    print(f"  HF Spaces Visit Summary")
+    print(f"{'='*40}")
+    for space, n in sorted(counts.items(), key=lambda x: -x[1]):
+        print(f"  {space:<35} {n:>6} visits")
+    print(f"  {'TOTAL':<35} {total:>6}")
+    print(f"{'='*40}\n")
+
+
+# ── Example integration ───────────────────────────────────────────────────────
+# In your app's main entry point (e.g. app.py), add:
+#
+#   from visit_counter import log_visit
+#   log_visit()   # <-- one line, runs at startup
+#
+# For Gradio specifically, put it just before gr.launch():
+#
+#   import gradio as gr
+#   from visit_counter import log_visit
+#   log_visit()
+#   demo.launch()
+#
+# For a Gradio "on load" event (fires each time UI loads in browser):
+#
+#   with gr.Blocks() as demo:
+#       demo.load(fn=log_visit)
+#   demo.launch()
+# ─────────────────────────────────────────────────────────────────────────────
+
+if __name__ == "__main__":
+    print_summary()

requirements.txt

@@ -10,3 +10,4 @@ joblib
 pytest
 tensorflow==2.9.1
 alfabet==0.2.2
+huggingface_hub