Improve Python bindings code quality (Codacy/Bandit cleanup) (#4084)

Author: robfrank

.github/workflows/test-python-bindings.yml

@@ -33,6 +33,29 @@ permissions:
   contents: read
 
 jobs:
+  bandit:
+    name: Bandit security scan (bindings/python)
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+
+      - name: Set up Python
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        with:
+          python-version: "3.12"
+
+      - name: Install Bandit
+        run: python -m pip install "bandit==1.9.4"
+
+      - name: Run Bandit on src and tests (must be clean)
+        working-directory: bindings/python
+        run: python -m bandit -c pyproject.toml -r src tests --severity-level low --confidence-level low
+
+      - name: Run Bandit on examples (must be clean at medium+/high-confidence)
+        working-directory: bindings/python
+        run: python -m bandit -c pyproject.toml -r examples --severity-level medium --confidence-level high
+
   # First job: Download ArcadeDB JARs (platform-agnostic)
   download-jars:
     name: Download ArcadeDB JARs

bindings/python/examples/11_vector_index_build.py

@@ -95,8 +95,10 @@ def get_docker_version() -> str | None:
 
 
 def fetch_json(url: str) -> dict:
+    if not url.startswith("https://"):
+        raise ValueError(f"Refusing to open non-HTTPS URL: {url!r}")
     req = Request(url, headers={"User-Agent": "arcadedb-bench"})
-    with urlopen(req, timeout=30) as response:
+    with urlopen(req, timeout=30) as response:  # nosec B310 - https-only
         payload = json.load(response)
     if not isinstance(payload, dict):
         raise RuntimeError(f"Expected JSON object from {url}")
@@ -870,7 +872,9 @@ def get_qdrant_version(client) -> str | None:
 
 
 def qdrant_project_name(db_path: Path) -> str:
-    digest = hashlib.sha1(str(db_path).encode("utf-8")).hexdigest()[:10]
+    digest = hashlib.sha1(
+        str(db_path).encode("utf-8"), usedforsecurity=False
+    ).hexdigest()[:10]
     return f"arcadb-qdrant-{digest}"
 
 
@@ -960,7 +964,9 @@ def wait_for_qdrant_ready(host: str, port: int, timeout_sec: int = 120) -> None:
     while True:
         for url in urls:
             try:
-                with urlopen(url, timeout=3) as response:
+                with urlopen(
+                    url, timeout=3
+                ) as response:  # nosec B310 - localhost health-check URL
                     if 200 <= int(response.status) < 500:
                         return
             except Exception:
@@ -1010,7 +1016,9 @@ def ensure_milvus_compose_file(compose_file: Path, release_tag: str) -> None:
             "https://github.com/milvus-io/milvus/releases/download/"
             f"{release_tag}/milvus-standalone-docker-compose.yml"
         )
-        urlretrieve(url, str(compose_file))
+        urlretrieve(
+            url, str(compose_file)
+        )  # nosec B310 - url is a hardcoded https://github.com URL
         raw = compose_file.read_text(encoding="utf-8")
 
     sanitized = re.sub(r"(?m)^\s*container_name:\s*.*\n", "", raw)
@@ -1024,7 +1032,9 @@ def ensure_milvus_compose_file(compose_file: Path, release_tag: str) -> None:
 
 
 def milvus_project_name(db_path: Path) -> str:
-    digest = hashlib.sha1(str(db_path).encode("utf-8")).hexdigest()[:10]
+    digest = hashlib.sha1(
+        str(db_path).encode("utf-8"), usedforsecurity=False
+    ).hexdigest()[:10]
     return f"arcadb-milvus-{digest}"
 
 

bindings/python/examples/12_vector_search.py

@@ -100,8 +100,10 @@ def get_docker_version() -> str | None:
 
 
 def fetch_json(url: str) -> dict:
+    if not url.startswith("https://"):
+        raise ValueError(f"Refusing to open non-HTTPS URL: {url!r}")
     req = Request(url, headers={"User-Agent": "arcadedb-bench"})
-    with urlopen(req, timeout=30) as response:
+    with urlopen(req, timeout=30) as response:  # nosec B310 - https-only
         payload = json.load(response)
     if not isinstance(payload, dict):
         raise RuntimeError(f"Expected JSON object from {url}")
@@ -914,7 +916,9 @@ def get_qdrant_version(client) -> str | None:
 
 
 def qdrant_project_name(db_path: Path) -> str:
-    digest = hashlib.sha1(str(db_path).encode("utf-8")).hexdigest()[:10]
+    digest = hashlib.sha1(
+        str(db_path).encode("utf-8"), usedforsecurity=False
+    ).hexdigest()[:10]
     return f"arcadb-qdrant-{digest}"
 
 
@@ -1004,7 +1008,9 @@ def wait_for_qdrant_ready(host: str, port: int, timeout_sec: int = 120) -> None:
     while True:
         for url in urls:
             try:
-                with urlopen(url, timeout=3) as response:
+                with urlopen(
+                    url, timeout=3
+                ) as response:  # nosec B310 - localhost health-check URL
                     if 200 <= int(response.status) < 500:
                         return
             except Exception:
@@ -1120,7 +1126,8 @@ def run_repeated_search(
             **run_stats,
             "run": run_idx + 1,
             "query_order_hash": hashlib.sha1(
-                ",".join(str(v) for v in run_qids).encode("utf-8")
+                ",".join(str(v) for v in run_qids).encode("utf-8"),
+                usedforsecurity=False,
             ).hexdigest(),
         }
         per_run_stats.append(run_stats)
@@ -1240,7 +1247,9 @@ def ensure_milvus_compose_file(compose_file: Path, release_tag: str) -> None:
             "https://github.com/milvus-io/milvus/releases/download/"
             f"{release_tag}/milvus-standalone-docker-compose.yml"
         )
-        urlretrieve(url, str(compose_file))
+        urlretrieve(
+            url, str(compose_file)
+        )  # nosec B310 - url is a hardcoded https://github.com URL
         raw = compose_file.read_text(encoding="utf-8")
 
     sanitized = re.sub(r"(?m)^\s*version\s*:\s*.*\n", "", raw)
@@ -1255,7 +1264,9 @@ def ensure_milvus_compose_file(compose_file: Path, release_tag: str) -> None:
 
 
 def milvus_project_name(db_path: Path) -> str:
-    digest = hashlib.sha1(str(db_path).encode("utf-8")).hexdigest()[:10]
+    digest = hashlib.sha1(
+        str(db_path).encode("utf-8"), usedforsecurity=False
+    ).hexdigest()[:10]
     return f"arcadb-milvus-{digest}"
 
 

bindings/python/examples/16_import_database_vs_transactional_graph_ingest.py

@@ -131,7 +131,10 @@ def edge_endpoints(edge_id: int, vertex_count: int) -> Tuple[int, int]:
 
 
 def build_rid_lookup_for_vertex_type(db, vertex_type: str) -> Dict[int, str]:
-    rows = db.query("sql", f"SELECT Id, @rid as rid FROM {vertex_type}").to_list()
+    rows = db.query(
+        "sql",
+        f"SELECT Id, @rid as rid FROM {vertex_type}",  # nosec B608 - vertex_type is a script constant
+    ).to_list()
     rid_lookup: Dict[int, str] = {}
     for row in rows:
         row_id = row.get("Id")
@@ -165,7 +168,12 @@ def collect_vertex_sample(
     db, vertex_type: str, vertex_id: int, props: List[ColumnDef]
 ) -> dict:
     row = query_one_or_none(
-        db.query("sql", f"SELECT FROM {vertex_type} WHERE Id = {vertex_id}")
+        db.query(
+            "sql",
+            # vertex_type is a constant from this script; vertex_id is bound as parameter.
+            f"SELECT FROM {vertex_type} WHERE Id = ?",  # nosec B608
+            vertex_id,
+        )
     )
     if row is None:
         return {"Id": vertex_id, "missing": True}
@@ -246,7 +254,8 @@ def collect_graph_signature(
         vertex_aggregate = query_one_or_none(
             db.query(
                 "sql",
-                f"SELECT {', '.join(vertex_aggregate_fields)} FROM {vertex_type}",
+                # vertex_aggregate_fields and vertex_type are script-local constants.
+                f"SELECT {', '.join(vertex_aggregate_fields)} FROM {vertex_type}",  # nosec B608
             )
         )
         edge_aggregate = query_one_or_none(

bindings/python/examples/17_timeseries_end_to_end.py

@@ -338,10 +338,13 @@ def main() -> int:
             for row in db.query(
                 "sql",
                 "SELECT FROM SensorReading "
-                f"WHERE ts BETWEEN {raw_window_start} AND {raw_window_end} "
-                f"AND sensor_id = '{focus_sensor.sensor_id}' "
-                f"AND building = '{focus_sensor.building}' "
+                "WHERE ts BETWEEN ? AND ? "
+                "AND sensor_id = ? AND building = ? "
                 "ORDER BY ts",
+                raw_window_start,
+                raw_window_end,
+                focus_sensor.sensor_id,
+                focus_sensor.building,
             )
         ]
         print_rows(

bindings/python/examples/20_graph_algorithms_route_planning.py

@@ -284,7 +284,8 @@ def insert_seed_data(db) -> None:
         for route in ROUTES:
             db.command(
                 "sql",
-                f"CREATE EDGE {route['edge_type']} "
+                # route['edge_type'] is a constant from the demo schema.
+                f"CREATE EDGE {route['edge_type']} "  # nosec B608
                 "FROM (SELECT FROM City WHERE code = ? LIMIT 1) "
                 "TO (SELECT FROM City WHERE code = ? LIMIT 1) "
                 "SET distance = ?, duration = ?, risk = ?, lane = ?",
@@ -755,7 +756,10 @@ def run_reopen_phase(db_path: Path) -> None:
             .get("count")
         )
         route_count = sum(
-            reopened_db.query("sql", f"SELECT count(*) AS count FROM {edge_type}")
+            reopened_db.query(
+                "sql",
+                f"SELECT count(*) AS count FROM {edge_type}",  # nosec B608 - edge_type is a script constant
+            )
             .first()
             .get("count")
             for edge_type in ("Road", "Rail", "Ferry")

bindings/python/examples/21_server_mode_http_access.py

@@ -164,8 +164,14 @@ def http_json_request(
         request_headers["Content-Type"] = "application/json"
 
     request = Request(url, data=data, headers=request_headers, method=method)
+    if not request.full_url.startswith(
+        ("http://localhost", "http://127.0.0.1", "https://")
+    ):
+        raise ValueError(f"Refusing to call unexpected URL: {request.full_url!r}")
     try:
-        with urlopen(request, timeout=timeout) as response:
+        with urlopen(
+            request, timeout=timeout
+        ) as response:  # nosec B310 - localhost or https
             body = response.read().decode("utf-8")
     except HTTPError as exc:
         detail = exc.read().decode("utf-8", errors="replace")

bindings/python/examples/22_graph_analytical_view_sql.py

@@ -539,6 +539,7 @@ def query_direct_neighbor_sample(
 
 
 def query_two_hop_summary(db, origin_code: str) -> dict:
+    # origin_code is a script-local constant from the demo dataset.
     result = db.query(
         "sql",
         f"""
@@ -550,7 +551,7 @@ def query_two_hop_summary(db, origin_code: str) -> dict:
                   {{type: City, as: dst}}
             RETURN DISTINCT dst.code AS code
         )
-        """,
+        """,  # nosec B608 - demo-data constants only
     )
     row = result.first()
     require(row is not None, "Expected a two-hop summary row")
@@ -567,7 +568,7 @@ def query_hub_inbound_count(db, hub_code: str) -> int:
                   {{type: City, as: hub, where: (code = '{hub_code}')}}
             RETURN src.code AS code
         )
-        """,
+        """,  # nosec B608 - demo-data constants only
     )
     row = result.first()
     require(row is not None, "Expected an inbound count row")
@@ -583,7 +584,7 @@ def query_region_sample(db, sample_limit: int) -> list[dict]:
         GROUP BY region
         ORDER BY region
         LIMIT {sample_limit}
-        """,
+        """,  # nosec B608 - sample_limit is a script integer constant
     )
     return rows_to_dicts(result, ["region", "city_count", "avg_demand"])
 

bindings/python/examples/download_data.py

@@ -113,6 +113,18 @@
     tqdm = None
 
 
+def _require_https(url: str) -> str:
+    """Reject non-HTTPS URLs before opening them.
+
+    Bandit B310 flags urlopen() because it permits file:// and custom schemes.
+    Examples download from a fixed list of HTTPS dataset URLs, so we enforce
+    that contract explicitly here.
+    """
+    if not url.startswith("https://"):
+        raise ValueError(f"Refusing to open non-HTTPS URL: {url!r}")
+    return url
+
+
 def ensure_clean_dir(path: Path, label: str) -> None:
     if path.exists():
         print(f"[CLEAN] Removing existing {label} directory: {path}")
@@ -188,8 +200,10 @@ def _download_with_python(
             f"[DOWNLOAD] Resuming {destination.name} from {_format_bytes(resume_from)}"
         )
 
-    request = urllib.request.Request(url, headers=headers)
-    with urllib.request.urlopen(request, timeout=60) as response:
+    request = urllib.request.Request(_require_https(url), headers=headers)
+    with urllib.request.urlopen(
+        request, timeout=60
+    ) as response:  # nosec B310 - https-only
         status = getattr(response, "status", response.getcode())
 
         if resume_from > 0 and status != 206:
@@ -649,7 +663,9 @@ def report_progress(block_num, block_size, total_size):
                 end="",
             )
 
-        urllib.request.urlretrieve(url, zip_path, reporthook=report_progress)
+        urllib.request.urlretrieve(
+            _require_https(url), zip_path, reporthook=report_progress
+        )  # nosec B310 - https-only
         print()  # New line after progress
         download_elapsed = time.time() - download_start
         print(f"[OK] Downloaded to: {zip_path} " f"({download_elapsed:.2f}s)")
@@ -1085,9 +1101,11 @@ def create_stackoverflow_large(
 
 
 def _iter_stackoverflow_rows(xml_path: Path, fields: list[str]):
-    import xml.etree.ElementTree as ET
+    import xml.etree.ElementTree as ET  # nosec B405 - parsing files we just downloaded over HTTPS and verified
 
-    context = ET.iterparse(xml_path, events=("start", "end"))
+    context = ET.iterparse(
+        xml_path, events=("start", "end")
+    )  # nosec B314 - input is a downloaded, checksum-verified file
     _, root = next(context)
     for event, elem in context:
         if event == "end" and elem.tag == "row":
@@ -1618,7 +1636,9 @@ def report_progress(block_num, block_size, total_size):
                 end="",
             )
 
-        urllib.request.urlretrieve(url, dbgen_zip, reporthook=report_progress)
+        urllib.request.urlretrieve(
+            _require_https(url), dbgen_zip, reporthook=report_progress
+        )  # nosec B310 - https-only
         print()
 
         extract_dir = data_dir / "tpch-dbgen-extract"
@@ -1701,7 +1721,13 @@ def download_ldbc_snb(scale_factor: int = 1) -> Path:
             "main/params-csv-merge-foreign.ini"
         )
         print("[DOWNLOAD] LDBC SNB params template")
-        template = urllib.request.urlopen(template_url).read().decode("utf-8")
+        template = (
+            urllib.request.urlopen(  # nosec B310 - https-only
+                _require_https(template_url)
+            )
+            .read()
+            .decode("utf-8")
+        )
         lines = []
         inserted = False
         for line in template.splitlines():
@@ -2139,7 +2165,7 @@ def verify_xml_nulls(extract_dir, sample_size=None):
     Returns:
         dict: Verification results
     """
-    import xml.etree.ElementTree as ET
+    import xml.etree.ElementTree as ET  # nosec B405 - parsing files we just downloaded over HTTPS and verified
 
     verification_start = time.time()
     results = {}
@@ -2163,7 +2189,9 @@ def verify_xml_nulls(extract_dir, sample_size=None):
         file_start = time.time()
 
         # Parse XML iteratively for large files
-        context = ET.iterparse(xml_path, events=("start", "end"))
+        context = ET.iterparse(
+            xml_path, events=("start", "end")
+        )  # nosec B314 - input is a downloaded, checksum-verified file
         _, root = next(context)  # Get root element
 
         all_attrs = set()