pyfunda 2.2.0__tar.gz → 2.2.1__tar.gz

{pyfunda-2.2.0 → pyfunda-2.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyfunda
-Version: 2.2.0
+Version: 2.2.1
 Summary: Python API for Funda.nl real estate listings
 Project-URL: Homepage, https://github.com/0xMH/pyfunda
 Project-URL: Repository, https://github.com/0xMH/pyfunda
@@ -567,6 +567,18 @@ if len(funda_prices) >= 2:
     print(f"\nPrice change: {change_pct:+.1f}%")
 ```
 
+## Disclaimer
+
+This is an unofficial library and is not affiliated with, authorized, maintained, sponsored, or endorsed by Funda or any of its affiliates. Use at your own risk.
+
+This library only accesses publicly available listing data through Funda's undocumented internal API. Using this library may violate Funda's Terms of Service. The authors are not responsible for any consequences of using this software.
+
+This project is intended for personal use, research, and educational purposes only.
+
+- The API is undocumented and may change or break at any time without notice.
+- Please use this library responsibly and avoid excessive requests that could burden Funda's infrastructure.
+- Scraped data may be subject to copyright and usage restrictions. Ensure your use complies with applicable laws.
+
 ## License
 
 AGPL-3.0

{pyfunda-2.2.0 → pyfunda-2.2.1}/README.md

@@ -545,6 +545,18 @@ if len(funda_prices) >= 2:
     print(f"\nPrice change: {change_pct:+.1f}%")
 ```
 
+## Disclaimer
+
+This is an unofficial library and is not affiliated with, authorized, maintained, sponsored, or endorsed by Funda or any of its affiliates. Use at your own risk.
+
+This library only accesses publicly available listing data through Funda's undocumented internal API. Using this library may violate Funda's Terms of Service. The authors are not responsible for any consequences of using this software.
+
+This project is intended for personal use, research, and educational purposes only.
+
+- The API is undocumented and may change or break at any time without notice.
+- Please use this library responsibly and avoid excessive requests that could burden Funda's infrastructure.
+- Scraped data may be subject to copyright and usage restrictions. Ensure your use complies with applicable laws.
+
 ## License
 
 AGPL-3.0

{pyfunda-2.2.0 → pyfunda-2.2.1}/funda/__init__.py

@@ -16,5 +16,5 @@ Example usage:
 from funda.funda import Funda, FundaAPI
 from funda.listing import Listing
 
-__version__ = "2.2.0"
+__version__ = "2.2.1"
 __all__ = ["Funda", "FundaAPI", "Listing", "__version__"]

{pyfunda-2.2.0 → pyfunda-2.2.1}/funda/funda.py

@@ -1,10 +1,12 @@
 """Main Funda API class."""
 
+import random
 import re
 import time
 from typing import Any
 
 from curl_cffi import requests
+from curl_cffi.const import CurlHttpVersion
 
 from funda.listing import Listing
 
@@ -16,17 +18,47 @@ API_LISTING_TINY = f"{API_BASE}/tinyId/{{tiny_id}}"
 API_SEARCH = "https://listing-search-wonen.funda.io/_msearch/template"
 API_WALTER = "https://api.walterliving.com/hunter/lookup"
 
-# Headers for mobile API
-HEADERS = {
-    "x-funda-app-platform": "android",
-    "content-type": "application/json",
-}
-
-SEARCH_HEADERS = {
-    "content-type": "application/json",
-    "accept": "application/json",
-    "referer": "https://www.funda.nl/",
-}
+FUNDA_JA3 = "771,4867-4865-4866-52393-52392-49195-49199-49196-49200-49161-49171-49162-49172-156-157-47-53,0-23-65281-10-11-35-13-51-45-43-21,29-23-24,0"
+
+
+
+def _make_headers(host: str, for_search: bool = False) -> list[tuple[str, str]]:
+    """Generate headers matching the Funda Android app."""
+    trace_id = str(random.randint(10**18, 10**19))
+    parent_id = hex(random.randint(10**15, 10**16))[2:]
+    tid = hex(int(time.time()))[2:] + "00000000"
+
+    headers = [
+        ("user-agent", "Dart/3.9 (dart:io)"),
+        ("x-datadog-sampling-priority", "0"),
+        ("x-datadog-origin", "rum"),
+        ("tracestate", f"dd=s:0;o:rum;p:{parent_id}"),
+        ("accept-encoding", "gzip"),
+        ("x-datadog-parent-id", trace_id),
+    ]
+
+    if for_search:
+        # Search endpoint uses referer and accept instead of x-funda-app-platform
+        headers.extend([
+            ("content-type", "application/json"),
+            ("referer", "https://www.funda.nl/"),
+            ("accept", "application/json"),
+        ])
+    else:
+        # Listing endpoint uses x-funda-app-platform
+        headers.extend([
+            ("x-funda-app-platform", "android"),
+            ("content-type", "application/json"),
+        ])
+
+    headers.extend([
+        ("traceparent", f"00-{tid}{trace_id[:16]}-{parent_id}-00"),
+        ("host", host),
+        ("x-datadog-tags", f"_dd.p.tid={tid}"),
+        ("x-datadog-trace-id", trace_id),
+    ])
+
+    return headers
 
 
 def _parse_area(value: str | None) -> int | None:
@@ -69,8 +101,7 @@ class Funda:
     def session(self) -> requests.Session:
         """Lazily create HTTP session."""
         if self._session is None:
-            self._session = requests.Session(impersonate="chrome")
-            self._session.headers.update(HEADERS)
+            self._session = requests.Session()
         return self._session
 
     def close(self) -> None:
@@ -111,17 +142,26 @@ class Funda:
 
         # Try tinyId endpoint first (8-9 digits), then globalId (7 digits)
         listing_id_str = str(listing_id)
+        host = "listing-detail-page.funda.io"
         if len(listing_id_str) >= 8:
             url = API_LISTING_TINY.format(tiny_id=listing_id_str)
         else:
             url = API_LISTING.format(listing_id=listing_id_str)
 
-        response = self.session.get(url, timeout=self.timeout)
+        headers = _make_headers(host)
+        response = self.session.get(
+            url, headers=headers, ja3=FUNDA_JA3,
+            http_version=CurlHttpVersion.V1_1, timeout=self.timeout
+        )
 
         # If tinyId fails, try as globalId
         if response.status_code == 404 and len(listing_id_str) >= 8:
             url = API_LISTING.format(listing_id=listing_id_str)
-            response = self.session.get(url, timeout=self.timeout)
+            headers = _make_headers(host)
+            response = self.session.get(
+                url, headers=headers, ja3=FUNDA_JA3,
+                http_version=CurlHttpVersion.V1_1, timeout=self.timeout
+            )
 
         if response.status_code != 200:
             raise LookupError(f"Listing {listing_id} not found")
@@ -259,11 +299,15 @@ class Funda:
         query = f"{index_line}\n{query_line}\n"
 
         # Retry on intermittent 400 errors from API
+        host = "listing-search-wonen.funda.io"
         for attempt in range(3):
+            headers = _make_headers(host, for_search=True)
             response = self.session.post(
                 API_SEARCH,
-                headers=SEARCH_HEADERS,
+                headers=headers,
                 data=query,
+                ja3=FUNDA_JA3,
+                http_version=CurlHttpVersion.V1_1,
                 timeout=self.timeout,
             )
             if response.status_code == 200:
@@ -458,11 +502,16 @@ class Funda:
         """
         consecutive_404s = 0
         current_id = since_id + 1
+        host = "listing-detail-page.funda.io"
 
         while consecutive_404s < max_consecutive_404s:
             url = API_LISTING.format(listing_id=current_id)
             try:
-                response = self.session.get(url, timeout=self.timeout)
+                headers = _make_headers(host)
+                response = self.session.get(
+                    url, headers=headers, ja3=FUNDA_JA3,
+                    http_version=CurlHttpVersion.V1_1, timeout=self.timeout
+                )
 
                 if response.status_code == 200:
                     consecutive_404s = 0
@@ -536,6 +585,7 @@ class Funda:
             json=payload,
             headers={"Accept": "application/json", "Content-Type": "application/json"},
             timeout=self.timeout,
+            http_version=CurlHttpVersion.V1_1,
         )
 
         if response.status_code != 200:

{pyfunda-2.2.0 → pyfunda-2.2.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "pyfunda"
-version = "2.2.0"
+version = "2.2.1"
 description = "Python API for Funda.nl real estate listings"
 readme = "README.md"
 license = "AGPL-3.0-or-later"

{pyfunda-2.2.0 → pyfunda-2.2.1}/test_all_flows.py

@@ -626,18 +626,18 @@ def test_session_lazy_loading():
     print("  Session lazy loading working correctly")
 
 
-@test("Session headers are set correctly")
+@test("Session is created correctly")
 def test_session_headers():
     from funda import Funda
 
     f = Funda()
     session = f.session
 
-    assert 'user-agent' in session.headers, "Should have user-agent header"
-    assert 'Dart' in session.headers['user-agent'], "User-agent should contain 'Dart'"
+    # Session exists (headers are generated per-request with Dart fingerprint)
+    assert session is not None, "Session should exist"
 
     f.close()
-    print("  Session headers set correctly")
+    print("  Session created correctly")
 
 
 @test("Custom timeout is respected")