From 13b3e142ccce222949c7b1cc018c98a2fc61bf4e Mon Sep 17 00:00:00 2001
From: Research Assistant <research@example.com>
Date: Mon, 27 Apr 2026 20:34:47 +0800
Subject: [PATCH] fix: OCR diagnostic L2 probe uses POST instead of GET
 (PaddleOCR rejects GET)

---
 paperforge/ocr_diagnostics.py | 12 +++---------
 tests/test_ocr_doctor.py      |  8 ++++----
 2 files changed, 7 insertions(+), 13 deletions(-)

diff --git a/paperforge/ocr_diagnostics.py b/paperforge/ocr_diagnostics.py
index 2824681e..313f3989 100644
--- a/paperforge/ocr_diagnostics.py
+++ b/paperforge/ocr_diagnostics.py
@@ -36,15 +36,16 @@ def ocr_doctor(config: dict[str, str] | None, live: bool = False) -> dict:
             "fix": "Set PADDLEOCR_API_TOKEN in .env or environment variables and re-run `paperforge ocr --diagnose`",
         }
 
-    # L2 — URL reachability
+    # L2 — URL reachability (use POST since most endpoints reject GET)
     job_url = os.environ.get(
         "PADDLEOCR_JOB_URL",
         "https://paddleocr.aistudio-app.com/api/v2/ocr/jobs",
     ).strip()
     try:
-        resp = requests.get(
+        resp = requests.post(
             job_url,
             headers={"Authorization": f"bearer {token}"},
+            json={"model": os.environ.get("PADDLEOCR_MODEL", "PaddleOCR-VL-1.5")},
             timeout=30,
         )
         if resp.status_code == 401:
@@ -62,13 +63,6 @@ def ocr_doctor(config: dict[str, str] | None, live: bool = False) -> dict:
                 "fix": "OCR provider is experiencing issues. Retry later with `paperforge ocr --diagnose`",
             }
         if resp.status_code != 200:
-            if resp.status_code == 405:
-                return {
-                    "level": 2,
-                    "passed": False,
-                    "error": "URL returned 405 Method Not Allowed",
-                    "fix": "PaddleOCR endpoint may require GET for probing but POST for OCR jobs. Check if your API endpoint supports both methods.",
-                }
             return {
                 "level": 2,
                 "passed": False,
diff --git a/tests/test_ocr_doctor.py b/tests/test_ocr_doctor.py
index a9e5f2f2..969ec441 100644
--- a/tests/test_ocr_doctor.py
+++ b/tests/test_ocr_doctor.py
@@ -25,13 +25,13 @@ def test_l1_missing_token():
 # L2 — URL reachability
 # ---------------------------------------------------------------------------
 def test_l2_bad_url():
-    """Mock GET returning 404 → level 2 failed."""
+    """Mock POST returning 404 → level 2 failed."""
     env = {"PADDLEOCR_API_TOKEN": "test-token"}
     mock_resp = MagicMock()
     mock_resp.status_code = 404
 
     with patch.dict(os.environ, env, clear=True):
-        with patch("paperforge.ocr_diagnostics.requests.get", return_value=mock_resp):
+        with patch("paperforge.ocr_diagnostics.requests.post", return_value=mock_resp):
             result = ocr_doctor(config=None, live=False)
 
     assert result["level"] == 2
@@ -41,13 +41,13 @@ def test_l2_bad_url():
 
 
 def test_l2_unauthorized():
-    """Mock GET returning 401 → level 2 failed with auth message."""
+    """Mock POST returning 401 → level 2 failed with auth message."""
     env = {"PADDLEOCR_API_TOKEN": "bad-token"}
     mock_resp = MagicMock()
     mock_resp.status_code = 401
 
     with patch.dict(os.environ, env, clear=True):
-        with patch("paperforge.ocr_diagnostics.requests.get", return_value=mock_resp):
+        with patch("paperforge.ocr_diagnostics.requests.post", return_value=mock_resp):
             result = ocr_doctor(config=None, live=False)
 
     assert result["level"] == 2