fix: Updated test case outputs and fuzzy assertion

tests/client_test.py

@@ -1,6 +1,7 @@
 import logging
 import os
 import unittest
+from difflib import SequenceMatcher, unified_diff
 from pathlib import Path
 
 import pytest
@@ -23,9 +24,7 @@ def test_get_usage_info(client):
         "subscription_plan",
         "today_page_count",
     ]
-    assert set(usage_info.keys()) == set(
-        expected_keys
-    ), f"usage_info {usage_info} does not contain the expected keys"
+    assert set(usage_info.keys()) == set(expected_keys), f"usage_info {usage_info} does not contain the expected keys"
 
 
 @pytest.mark.parametrize(
@@ -56,7 +55,21 @@ def test_whisper(client, data_dir, processing_mode, output_mode, input_file):
 
     assert isinstance(response, dict)
     assert response["status_code"] == 200
-    assert response["extracted_text"] == exp
+
+    # For text based processing, perform a strict match
+    if processing_mode == "text" and output_mode == "text":
+        assert response["extracted_text"] == exp
+    # For OCR based processing, perform a fuzzy match
+    else:
+        extracted_text = response["extracted_text"]
+        similarity = SequenceMatcher(None, extracted_text, exp).ratio()
+        threshold = 0.97
+
+        if similarity < threshold:
+            diff = "\n".join(
+                unified_diff(exp.splitlines(), extracted_text.splitlines(), fromfile="Expected", tofile="Extracted")
+            )
+            pytest.fail(f"Texts are not similar enough: {similarity * 100:.2f}% similarity. Diff:\n{diff}")
 
 
 # TODO: Review and port to pytest based tests
@@ -78,9 +91,7 @@ def test_whisper(self):
     # @unittest.skip("Skipping test_whisper")
     def test_whisper_stream(self):
         client = LLMWhispererClient()
-        download_url = (
-            "https://storage.googleapis.com/pandora-static/samples/bill.jpg.pdf"
-        )
+        download_url = "https://storage.googleapis.com/pandora-static/samples/bill.jpg.pdf"
         # Create a stream of download_url and pass it to whisper
         response_download = requests.get(download_url, stream=True)
         response_download.raise_for_status()
@@ -95,18 +106,14 @@ def test_whisper_stream(self):
     @unittest.skip("Skipping test_whisper_status")
     def test_whisper_status(self):
         client = LLMWhispererClient()
-        response = client.whisper_status(
-            whisper_hash="7cfa5cbb|5f1d285a7cf18d203de7af1a1abb0a3a"
-        )
+        response = client.whisper_status(whisper_hash="7cfa5cbb|5f1d285a7cf18d203de7af1a1abb0a3a")
         logger.info(response)
         self.assertIsInstance(response, dict)
 
     @unittest.skip("Skipping test_whisper_retrieve")
     def test_whisper_retrieve(self):
         client = LLMWhispererClient()
-        response = client.whisper_retrieve(
-            whisper_hash="7cfa5cbb|5f1d285a7cf18d203de7af1a1abb0a3a"
-        )
+        response = client.whisper_retrieve(whisper_hash="7cfa5cbb|5f1d285a7cf18d203de7af1a1abb0a3a")
         logger.info(response)
         self.assertIsInstance(response, dict)
 

tests/test_data/expected/handwritten-form.ocr.line-printer.txt

@@ -14,7 +14,7 @@ How did you hear about us?
 [ ] This mailer 
 [X] Other (Explain) SAW THE SIGN WHEN THE LOCATION 
 
-            WAS    BEING BUILT 
+            WAS    BEING   BUILT 
 
 By signing, I agree to receive all communications from acme, inc. 
 

tests/test_data/expected/handwritten-form.text.text.txt

@@ -2,12 +2,20 @@ Name
 STEPHEN YOUNG
 Address
 123 MAIN ST.
-Are you: :unselected: Married :selected: Single
-How did you hear about us? :unselected: Search Ad :unselected: Facebook :unselected: :selected: X (formerly Twitter) :unselected: This mailer :selected: Other (Explain) ; SAW THE SIGN WHEN THE LOCATION WAS BEING BUILT
+Are you:
+[ ] Married
+[X] Single
+How did you hear about us?
+[ ] Search Ad
+[ ] Facebook
+[ ] [X] X (formerly Twitter)
+[ ] This mailer
+[X] Other (Explain) i SAW THE SIGN WHEN THE LOCATION WAS BEING BUILT
 By signing, I agree to receive all communications from acme, inc.
 Signature
 Date
 10/15/23
 
 
+
 <<<

tests/test_data/expected/restaurant_invoice_photo.ocr.line-printer.txt

@@ -11,30 +11,27 @@
 
  Type : Table 
 
- Table Number: 2 
+ Table Number : 2 
 
  Bill No .: T2 -- 126653 
-
  Date:2023-05-31 23:16:50 
  Kots:63 
 
  Item      Qty Amt 
 
  Jack The 
-
  Ripper   1     400.00 
-Plain Fries + 
-Coke 300 ML 1   130.00 
+ Plain Fries + 
+ Coke 300 ML 1 130.00 
 
  Total Qty: 2 
- SubTotal :     530.00 
+ SubTotal:      530.00 
 
 GST@5%          26.50 
-
-   CGST @2.5% 13.25 
+   CGST @2.5%. 13.25 
    SGST @2.5% 13.25 
 
-Round Off:       0.50 
+Round Off :      0.50 
 
 Total Invoice Value: 557 
 

tests/test_data/expected/restaurant_invoice_photo.ocr.text.txt

@@ -1,5 +1,5 @@
 BURGER SEIGNEUR No. 35, 80 feet road, HAL 3rd Stage, Indiranagar, Bangalore GST: 29AAHFL9534H1ZV
-Order Number : T2- 57 Type : Table Table Number: 2
+Order Number : T2- 57 Type : Table Table Number : 2
 Bill No .: T2 -- 126653 Date:2023-05-31 23:16:50 Kots:63
 Item
 Qty
@@ -14,20 +14,23 @@ Coke 300 ML
 130.00
 Total Qty:
 2
-SubTotal :
+SubTotal:
 530.00
 GST@5%
 26.50
-CGST @2.5%
+CGST @2.5%.
 13.25
 SGST @2.5%
 13.25
-Round Off:
+Round Off :
 0.50
-Total Invoice Value:
+Total Invoice
+Value:
 557
-PAY: 557 Thank you, visit again! :selected:
+PAY: 557
+Thank you, visit again!
 Powered by - POSIST
 
 
+
 <<<

tests/test_data/expected/restaurant_invoice_photo.text.line-printer.txt

@@ -11,30 +11,27 @@
 
  Type : Table 
 
- Table Number: 2 
+ Table Number : 2 
 
  Bill No .: T2 -- 126653 
-
  Date:2023-05-31 23:16:50 
  Kots:63 
 
  Item      Qty Amt 
 
  Jack The 
-
  Ripper   1     400.00 
-Plain Fries + 
-Coke 300 ML 1   130.00 
+ Plain Fries + 
+ Coke 300 ML 1 130.00 
 
  Total Qty: 2 
- SubTotal :     530.00 
+ SubTotal:      530.00 
 
 GST@5%          26.50 
-
-   CGST @2.5% 13.25 
+   CGST @2.5%. 13.25 
    SGST @2.5% 13.25 
 
-Round Off:       0.50 
+Round Off :      0.50 
 
 Total Invoice Value: 557