Page de statistiques

tools/stats/top_requests.py

@@ -0,0 +1,107 @@
+#!/usr/bin/env python3
+
+import sys
+import json
+import os
+import urllib.request
+
+CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
+TOOLS_DIR = os.path.abspath(os.path.join(CURRENT_DIR, os.pardir))
+ROOT_DIR = os.path.abspath(os.path.join(TOOLS_DIR, os.pardir))
+if CURRENT_DIR not in sys.path:
+    sys.path.append(CURRENT_DIR)
+if TOOLS_DIR not in sys.path:
+    sys.path.append(TOOLS_DIR)
+
+
+def load_config():
+    cfg_path = os.path.join(ROOT_DIR, "tools", "config.json")
+    try:
+        with open(cfg_path, "r", encoding="utf-8") as handle:
+            return json.load(handle)
+    except Exception:
+        return {}
+
+
+def fetch_goaccess(url, timeout=10):
+    with urllib.request.urlopen(url, timeout=timeout) as resp:
+        data = resp.read().decode("utf-8")
+        return json.loads(data)
+
+
+def crawler_ratios(data):
+    browsers = (data.get("browsers") or {}).get("data") or []
+    crawler = next((entry for entry in browsers if entry.get("data") == "Crawlers"), None)
+    if not crawler:
+        return {"hits": 0.0, "visitors": 0.0}
+
+    def total(field):
+        return sum((entry.get(field, {}) or {}).get("count", 0) for entry in browsers)
+
+    total_hits = total("hits")
+    total_visitors = total("visitors")
+    return {
+        "hits": min(1.0, (crawler.get("hits", {}) or {}).get("count", 0) / total_hits) if total_hits else 0.0,
+        "visitors": min(1.0, (crawler.get("visitors", {}) or {}).get("count", 0) / total_visitors)
+        if total_visitors
+        else 0.0,
+    }
+
+
+def adjust(value, ratio):
+    return max(0, round(value * (1 - ratio)))
+
+
+def main():
+    try:
+        payload = json.load(sys.stdin)
+    except Exception as exc:  # noqa: BLE001
+        print(f"Failed to read JSON: {exc}", file=sys.stderr)
+        sys.exit(1)
+
+    output_path = payload.get("outputPath")
+    public_path = payload.get("publicPath")
+    url = payload.get("stat", {}).get("url")
+
+    cfg = load_config()
+    goaccess_url = url or (cfg.get("goaccess") or {}).get("url") or ""
+
+    try:
+        data = fetch_goaccess(goaccess_url)
+    except Exception as exc:  # noqa: BLE001
+        print(f"Failed to fetch GoAccess JSON: {exc}", file=sys.stderr)
+        sys.exit(1)
+
+    ratios = crawler_ratios(data)
+
+    reqs = (data.get("requests") or {}).get("data") or []
+    # entries have .data = path, hits.count, visitors.count ?
+    cleaned = []
+    for entry in reqs:
+        path = entry.get("data") or ""
+        hits = (entry.get("hits") or {}).get("count", 0)
+        if not path or hits <= 0:
+            continue
+        cleaned.append((path, adjust(hits, ratios["hits"])))
+
+    cleaned.sort(key=lambda item: item[1], reverse=True)
+    top = cleaned[:10]
+
+    labels = [item[0] for item in top]
+    values = [item[1] for item in top]
+
+    try:
+        from render_stats_charts import render_top_requests, setup_rcparams
+    except ImportError as exc:  # noqa: BLE001
+        print(f"Failed to import renderer: {exc}", file=sys.stderr)
+        sys.exit(1)
+
+    setup_rcparams()
+    render_top_requests({"labels": labels, "values": values, "title": "Top 10 requêtes (hors crawlers)"}, output_path)
+
+    json.dump({"image": public_path}, sys.stdout)
+    sys.stdout.flush()
+
+
+if __name__ == "__main__":
+    main()