mediacatch 0.2.2__tar.gz → 0.2.4__tar.gz

{mediacatch-0.2.2 → mediacatch-0.2.4}/PKG-INFO

@@ -1,11 +1,9 @@
 Metadata-Version: 2.3
 Name: mediacatch
-Version: 0.2.2
+Version: 0.2.4
 Summary: Python package for easily interfacing with MediaCatch APIs and data formats.
-Project-URL: Documentation, https://github.com/mediacatch/mediacatch#readme
-Project-URL: Issues, https://github.com/mediacatch/mediacatch/issues
-Project-URL: Source, https://github.com/mediacatch/mediacatch
-Author-email: Mathias Nielsen <mathias@mediacatch.io>, Frederik Haarslev <frederik@mediacatch.io>
+Project-URL: Documentation, https://api.mediacatch.io/mediacatch/docs/
+Author-email: Mathias Nielsen <mathias@mediacatch.io>
 License-Expression: MIT
 Classifier: Framework :: FastAPI
 Classifier: Intended Audience :: Developers
@@ -49,13 +47,25 @@ Requires-Dist: pytest-cov; extra == 'test'
 Requires-Dist: ruff; extra == 'test'
 Description-Content-Type: text/markdown
 
+<div align="center">
+
+[<img src="docs/assets/mc_logo.svg" alt="mc-logo" width=400>](https://api.mediacatch.io/mediacatch/docs/)<br>
+
 [![Unit Tests](https://github.com/mediacatch/mediacatch/actions/workflows/test.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/test.yaml)
 [![codecov](https://codecov.io/gh/mediacatch/mediacatch/graph/badge.svg?token=D3bBQ64Cs9)](https://codecov.io/gh/mediacatch/mediacatch)
 [![Quality Check](https://github.com/mediacatch/mediacatch/actions/workflows/check.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/check.yaml)
 [![Publish](https://github.com/mediacatch/mediacatch/actions/workflows/publish.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/publish.yaml)
-[![Documentation](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml)
+[![Documentation](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml)<br>
+
+<h4 align="center">
+    <p>
+        <a href="#requirements">Requirements</a> |
+        <a href="#installation">Installation</a> |
+        <a href="#getting-started">Getting Started</a> |
+    <p>
+</h4>
 
-# MediaCatch
+</div>
 
 Python package for easily interfacing with MediaCatch APIs and data formats. Documentation can be found at [api.mediacatch.io/mediacatch/docs/](https://api.mediacatch.io/mediacatch/docs/).
 

{mediacatch-0.2.2 → mediacatch-0.2.4}/README.md

@@ -1,10 +1,22 @@
+<div align="center">
+
+[<img src="docs/assets/mc_logo.svg" alt="mc-logo" width=400>](https://api.mediacatch.io/mediacatch/docs/)<br>
+
 [![Unit Tests](https://github.com/mediacatch/mediacatch/actions/workflows/test.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/test.yaml)
 [![codecov](https://codecov.io/gh/mediacatch/mediacatch/graph/badge.svg?token=D3bBQ64Cs9)](https://codecov.io/gh/mediacatch/mediacatch)
 [![Quality Check](https://github.com/mediacatch/mediacatch/actions/workflows/check.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/check.yaml)
 [![Publish](https://github.com/mediacatch/mediacatch/actions/workflows/publish.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/publish.yaml)
-[![Documentation](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml)
+[![Documentation](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml/badge.svg)](https://github.com/mediacatch/mediacatch/actions/workflows/documentation.yaml)<br>
+
+<h4 align="center">
+    <p>
+        <a href="#requirements">Requirements</a> |
+        <a href="#installation">Installation</a> |
+        <a href="#getting-started">Getting Started</a> |
+    <p>
+</h4>
 
-# MediaCatch
+</div>
 
 Python package for easily interfacing with MediaCatch APIs and data formats. Documentation can be found at [api.mediacatch.io/mediacatch/docs/](https://api.mediacatch.io/mediacatch/docs/).
 

mediacatch-0.2.4/docs/assets/mc_logo.svg

@@ -0,0 +1,21 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg width="259px" height="32px" viewBox="0 0 259 32" version="1.1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
+    <title>mc_logo</title>
+    <g id="Social-media" stroke="none" stroke-width="1" fill="none" fill-rule="evenodd">
+        <g id="mc_logo" fill="#000000" fill-rule="nonzero">
+            <path d="M5.2261008,0.0132134395 L29.8682986,2.59855825 C31.1757299,2.73575455 32.1242044,3.90494078 31.986763,5.21004021 L29.3900862,29.8720617 C29.2526447,31.1771612 28.0813695,32.123944 26.7739382,31.9867952 L2.13174035,29.4014029 C0.824309029,29.2642541 -0.124165435,28.0950679 0.0132283961,26.7899684 L2.60995277,2.12789937 C2.74739421,0.822799936 3.91866948,-0.123982864 5.2261008,0.0132134395 Z M24.4055811,6.33082069 C24.2693398,5.47002125 23.4594639,4.88305274 22.5966754,5.01976749 L6.33509989,7.59686183 C5.47231134,7.73357659 4.88334655,8.5422238 5.01958783,9.40302324 L7.5943965,25.6692099 C7.73066834,26.5300012 8.54054424,27.1169697 9.40332462,26.9802244 L25.6649083,24.4031606 C26.5276887,24.2664153 27.1166534,23.4577681 26.9804122,22.5969687 L24.4055811,6.33082069 Z" id="Shape"></path>
+            <g id="MEDIACATCH" transform="translate(45, 6)">
+                <polygon id="Path" points="19.620064 1 12.0384205 10.5142857 4.45677695 1 0 1 0 19 3.9701174 19 3.9701174 6.47714286 11.8335112 16.12 12.0384205 16.12 19.9018143 6.47714286 19.9018143 19 24 19 24 1"></polygon>
+                <polygon id="Path" points="33.1900139 15.5542857 33.1900139 11.5942857 45.3384189 11.5942857 45.3384189 8.14857143 33.1900139 8.14857143 33.1900139 4.44571429 47.7364771 4.44571429 47.7364771 1 29 1 29 19 48 19 48 15.5542857"></polygon>
+                <path d="M60.7043478,1 L51,1 L51,19 L60.7043478,19 C67.5913043,19 72,15.6057143 72,10 C72,4.39428571 67.5913043,1 60.7043478,1 Z M60.3913043,15.5028571 L55.173913,15.5028571 L55.173913,4.49714286 L60.3913043,4.49714286 C64.773913,4.49714286 67.7478261,6.29714286 67.7478261,10 C67.7478261,13.7028571 64.773913,15.5028571 60.3913043,15.5028571 Z" id="Shape"></path>
+                <polygon id="Path" points="80 19 80 1 76 1 76 19"></polygon>
+                <path d="M102.472192,19 L107,19 L97.2900763,1 L92.7099237,1 L83,19 L87.4754635,19 L89.438386,15.1942857 L100.509269,15.1942857 L102.472192,19 Z M91.1657579,11.8514286 L94.9607415,4.52285714 L98.7557252,11.8514286 L91.1657579,11.8514286 Z" id="Shape"></path>
+                <path d="M118.846154,17.0026954 C113.641026,17.0026954 110.205128,14.4420485 110.205128,9.5 C110.205128,4.55795148 113.641026,1.99730458 118.846154,1.99730458 C121.974359,1.99730458 124.538462,3.04716981 125.948718,5.42857143 L128,4.40431267 C126.307692,1.74123989 123.179487,0 118.871795,0 C112.102564,0 108,3.68733154 108,9.5 C108,15.3126685 112.102564,19 118.794872,19 C123.179487,19 126.307692,17.2843666 128,14.5956873 L125.948718,13.5714286 C124.564103,15.9784367 121.974359,17.0026954 118.846154,17.0026954 Z" id="Path"></path>
+                <path d="M149.471065,19 L152,19 L141.751157,1 L139.248843,1 L129,19 L131.475694,19 L134.190972,14.1657143 L146.755787,14.1657143 L149.471065,19 Z M135.282407,12.2114286 L140.47338,2.98 L145.664352,12.2114286 L135.282407,12.2114286 Z" id="Shape"></path>
+                <polygon id="Path" points="169 1 150 1 150 2.95428571 158.416107 2.95428571 158.416107 19 160.558389 19 160.558389 2.95428571 169 2.95428571"></polygon>
+                <path d="M181.846154,17.0026954 C176.641026,17.0026954 173.205128,14.4420485 173.205128,9.5 C173.205128,4.55795148 176.641026,1.99730458 181.846154,1.99730458 C184.974359,1.99730458 187.538462,3.04716981 188.948718,5.42857143 L191,4.40431267 C189.307692,1.74123989 186.179487,0 181.871795,0 C175.102564,0 171,3.68733154 171,9.5 C171,15.3126685 175.102564,19 181.794872,19 C186.179487,19 189.307692,17.2843666 191,14.5956873 L188.948718,13.5714286 C187.564103,15.9784367 184.974359,17.0026954 181.846154,17.0026954 Z" id="Path"></path>
+                <polygon id="Path" points="211.854839 1 211.854839 8.84285714 197.145161 8.84285714 197.145161 1 195 1 195 19 197.145161 19 197.145161 10.9 211.854839 10.9 211.854839 19 214 19 214 1"></polygon>
+            </g>
+        </g>
+    </g>
+</svg>

{mediacatch-0.2.2 → mediacatch-0.2.4}/mediacatch/__init__.py

@@ -14,4 +14,4 @@ logging.basicConfig(
 load_dotenv()
 mediacatch_api_key = os.environ['MEDIACATCH_API_KEY']
 
-__version__ = '0.2.2'
+__version__ = '0.2.4'

{mediacatch-0.2.2 → mediacatch-0.2.4}/mediacatch/commands/vision.py

@@ -26,6 +26,8 @@ def vision_cli_factory(args):
         max_height_width_ratio=args.max_height_width_ratio,
         get_detection_histogram=args.get_detection_histogram,
         detection_histogram_bins=args.detection_histogram_bins,
+        max_height_difference_ratio=args.max_height_difference_ratio,
+        max_horizontal_distance_ratio=args.max_horizontal_distance_ratio,
         only_upload=args.only_upload,
     )
 
@@ -115,6 +117,18 @@ class VisionCLI(BaseCLICommand):
             default=None,
             help='Number of bins for histogram calculation',
         )
+        vision_parser.add_argument(
+            '--max-height-difference-ratio',
+            type=float,
+            default=None,
+            help='Determine the maximum allowed difference in height between two text boxes for them to be merged',
+        )
+        vision_parser.add_argument(
+            '--max-horizontal-distance-ratio',
+            type=float,
+            default=None,
+            help='Determine if two boxes are close enough horizontally to be considered part of the same text line',
+        )
         vision_parser.add_argument(
             '--only-upload',
             action='store_true',
@@ -138,6 +152,8 @@ class VisionCLI(BaseCLICommand):
         max_height_width_ratio: float | None = None,
         get_detection_histogram: bool | None = None,
         detection_histogram_bins: int | None = None,
+        max_height_difference_ratio: float | None = None,
+        max_horizontal_distance_ratio: float | None = None,
         only_upload: bool = False,
     ) -> None:
         self.file_path = file_path
@@ -154,6 +170,8 @@ class VisionCLI(BaseCLICommand):
         self.max_height_width_ratio = max_height_width_ratio
         self.get_detection_histogram = get_detection_histogram
         self.detection_histogram_bins = detection_histogram_bins
+        self.max_height_difference_ratio = max_height_difference_ratio
+        self.max_horizontal_distance_ratio = max_horizontal_distance_ratio
         self.only_upload = only_upload
 
     def run(self) -> None:
@@ -172,6 +190,8 @@ class VisionCLI(BaseCLICommand):
             max_height_width_ratio=self.max_height_width_ratio,
             get_detection_histogram=self.get_detection_histogram,
             detection_histogram_bins=self.detection_histogram_bins,
+            max_height_difference_ratio=self.max_height_difference_ratio,
+            max_horizontal_distance_ratio=self.max_horizontal_distance_ratio,
         )
         if self.only_upload:
             logger.info(f'Find result at https://api.mediacatch.io/vision/result/{file_id}')

{mediacatch-0.2.2 → mediacatch-0.2.4}/mediacatch/speech/result.py

@@ -52,7 +52,8 @@ def wait_for_result(
             return result
 
         except (requests.RequestException, json.JSONDecodeError):
-            time.sleep(delay)
+            if verbose:
+                logger.error('Error occurred while waiting for JSON response')
 
         except Exception as e:
             if verbose:
@@ -60,6 +61,8 @@ def wait_for_result(
 
             raise MediacatchError(f'Failed to get result from {result_url}: {e}')
 
+        time.sleep(delay)
+
     if verbose:
         logger.error(f'Timeout waiting for result from {result_url}, give up')
 

{mediacatch-0.2.2 → mediacatch-0.2.4}/mediacatch/vision/result.py

@@ -45,7 +45,23 @@ def wait_for_result(
 
                 raise MediacatchAPIError(response.status_code, response.json())
 
+            if response.status_code == 102:
+                if verbose:
+                    logger.info(f'Waiting for result from {result_url}')
+
+                time.sleep(delay)
+                continue
+            elif response.status_code == 504:
+                time.sleep(delay)
+                continue
+
             response.raise_for_status()
+            if response.status_code == 204:
+                if verbose:
+                    logger.info(f'No results found for {file_id}')
+
+                return {}
+
             result = response.json()
             elapsed_time = time.time() - start_time
             if verbose:
@@ -54,7 +70,8 @@ def wait_for_result(
             return result
 
         except (requests.RequestException, json.JSONDecodeError):
-            time.sleep(delay)
+            if verbose:
+                logger.error('Error occurred while waiting for JSON response')
 
         except Exception as e:
             if verbose:
@@ -62,6 +79,8 @@ def wait_for_result(
 
             raise MediacatchError(f'Failed to get result from {result_url}: {e}')
 
+        time.sleep(delay)
+
     if verbose:
         logger.error(f'Timeout waiting for result from {result_url}, give up')
 
@@ -69,6 +88,6 @@ def wait_for_result(
 
 
 if __name__ == '__main__':
-    file_id = '667bc663c8a5993596bf2803'
+    file_id = '66e6fb29b775ebfd11fb40f7'
     result = wait_for_result(file_id)
     print(result)

{mediacatch-0.2.2 → mediacatch-0.2.4}/mediacatch/vision/upload.py

@@ -34,6 +34,8 @@ def upload(
     max_height_width_ratio: float | None = None,
     get_detection_histogram: bool | None = None,
     detection_histogram_bins: int | None = None,
+    max_height_difference_ratio: float | None = None,
+    max_horizontal_distance_ratio: float | None = None,
     get_frame_index: bool | None = None,
     get_bbox: bool | None = None,
     face_recognition: bool | None = None,
@@ -62,6 +64,8 @@ def upload(
         max_height_width_ratio (float, optional): Discard detection if height/width ratio is greater than this value. Defaults to 2.0.
         get_detection_histogram (bool, optional): If true, get histogram of detection. Defaults to False.
         detection_histogram_bins (int, optional): Number of bins for histogram calculation. Defaults to 8.
+        max_height_difference_ratio (float, optional): Determine the maximum allowed difference in height between two text boxes for them to be merged. Defaults to 0.5.
+        max_horizontal_distance_ratio (float, optional): Determine if two boxes are close enough horizontally to be considered part of the same text line. Defaults to 0.9.
         get_frame_index (bool, optional): If true, get frame index. Defaults to None.
         get_bbox (bool, optional): If true, get bounding box. Defaults to None.
         face_recognition (bool, optional): If true, run face recognition. Defaults to None.
@@ -102,6 +106,10 @@ def upload(
         extra['get_detection_histogram'] = get_detection_histogram
     if detection_histogram_bins is not None:
         extra['detection_histogram_bins'] = detection_histogram_bins
+    if max_height_difference_ratio is not None:
+        extra['max_height_difference_ratio'] = max_height_difference_ratio
+    if max_horizontal_distance_ratio is not None:
+        extra['max_horizontal_distance_ratio'] = max_horizontal_distance_ratio
     if get_frame_index is not None:
         extra['frame'] = get_frame_index
     if get_bbox is not None:

{mediacatch-0.2.2 → mediacatch-0.2.4}/pyproject.toml

@@ -3,9 +3,7 @@ requires = ["hatchling"]
 build-backend = "hatchling.build"
 
 [project.urls]
-Documentation = "https://github.com/mediacatch/mediacatch#readme"
-Issues = "https://github.com/mediacatch/mediacatch/issues"
-Source = "https://github.com/mediacatch/mediacatch"
+Documentation = "https://api.mediacatch.io/mediacatch/docs/"
 
 [tool.hatch.version]
 path = "mediacatch/__init__.py"
@@ -17,8 +15,7 @@ readme = "README.md"
 requires-python = ">=3.10"
 license = "MIT"
 authors = [
-    { name = "Mathias Nielsen", email = "mathias@mediacatch.io" },
-    { name = "Frederik Haarslev", email = "frederik@mediacatch.io" },
+    { name = "Mathias Nielsen", email = "mathias@mediacatch.io" }
 ]
 classifiers = [
   "Topic :: Internet",