pycaption 2.2.16__tar.gz → 2.2.18__tar.gz

{pycaption-2.2.16 → pycaption-2.2.18}/MANIFEST.in

@@ -1,3 +1,2 @@
-include pycaption/english.pickle
 recursive-include tests *
 include README.rst

{pycaption-2.2.16/pycaption.egg-info → pycaption-2.2.18}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: pycaption
-Version: 2.2.16
+Version: 2.2.18
 Summary: Closed caption converter
 Author: Joe Norton
 Author-email: joey@nortoncrew.com
@@ -29,11 +29,12 @@ Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-lazy-fixture; extra == "dev"
 Provides-Extra: transcript
-Requires-Dist: nltk==3.8.0; extra == "transcript"
+Requires-Dist: nltk==3.9.1; extra == "transcript"
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier
 Dynamic: description
+Dynamic: license-file
 Dynamic: project-url
 Dynamic: provides-extra
 Dynamic: requires-dist

pycaption-2.2.18/pycaption/transcript.py

@@ -0,0 +1,33 @@
+from pycaption.base import BaseWriter, CaptionNode
+
+
+class TranscriptWriter(BaseWriter):
+    def __init__(self, *args, **kw):
+        try:
+            from nltk import PunktSentenceTokenizer
+
+            self.tokenizer = PunktSentenceTokenizer()
+        except ModuleNotFoundError as exc:
+            raise ModuleNotFoundError(
+                "Missing Dependency: You must install nltk"
+            ) from exc
+
+    def write(self, captions):
+        transcripts = []
+
+        for lang in captions.get_languages():
+            lang_transcript = ""
+
+            for caption in captions.get_captions(lang):
+                lang_transcript = self._strip_text(caption.nodes, lang_transcript)
+
+            lang_transcript = "\n".join(self.tokenizer.tokenize(lang_transcript))
+            transcripts.append(lang_transcript)
+
+        return "\n".join(transcripts)
+
+    def _strip_text(self, elements, lang_transcript):
+        return " ".join(
+            [lang_transcript]
+            + [el.content for el in elements if el.type_ == CaptionNode.TEXT]
+        )

{pycaption-2.2.16 → pycaption-2.2.18/pycaption.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: pycaption
-Version: 2.2.16
+Version: 2.2.18
 Summary: Closed caption converter
 Author: Joe Norton
 Author-email: joey@nortoncrew.com
@@ -29,11 +29,12 @@ Provides-Extra: dev
 Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-lazy-fixture; extra == "dev"
 Provides-Extra: transcript
-Requires-Dist: nltk==3.8.0; extra == "transcript"
+Requires-Dist: nltk==3.9.1; extra == "transcript"
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier
 Dynamic: description
+Dynamic: license-file
 Dynamic: project-url
 Dynamic: provides-extra
 Dynamic: requires-dist

{pycaption-2.2.16 → pycaption-2.2.18}/pycaption.egg-info/SOURCES.txt

@@ -4,7 +4,6 @@ README.rst
 setup.py
 pycaption/__init__.py
 pycaption/base.py
-pycaption/english.pickle
 pycaption/exceptions.py
 pycaption/geometry.py
 pycaption/microdvd.py

{pycaption-2.2.16 → pycaption-2.2.18}/pycaption.egg-info/requires.txt

@@ -7,4 +7,4 @@ pytest
 pytest-lazy-fixture
 
 [transcript]
-nltk==3.8.0
+nltk==3.9.1

{pycaption-2.2.16 → pycaption-2.2.18}/setup.py

@@ -16,11 +16,11 @@ dependencies = [
 
 dev_dependencies = ["pytest", "pytest-lazy-fixture"]
 
-transcript_dependencies = ["nltk==3.8.0"]
+transcript_dependencies = ["nltk==3.9.1"]
 
 setup(
     name="pycaption",
-    version="2.2.16",
+    version="2.2.18",
     description="Closed caption converter",
     long_description=open(README_PATH).read(),
     author="Joe Norton",