modify dependency for docx

app.py

@@ -112,84 +112,6 @@ def get_mp3(text: str, voice: str, api_key: str = None) -> bytes:
             for chunk in response.iter_bytes():
                 file.write(chunk)
             return file.getvalue()
-# def get_ppt_text(filename):
-#     prs = Presentation(filename)
-#     full_text = []
-#     for slide in prs.slides:
-#         for shape in slide.shapes:
-#             if hasattr(shape, "text"):
-#                 full_text.append(shape.text)
-#     return '\n'.join(full_text)
-
-
-# def get_doc_text(filename):
-#     doc = Document(filename)
-#     full_text = [para.text for para in doc.paragraphs]
-#     return '\n'.join(full_text)
-
-
-# def get_page_text(filename, page_num):
-#     with pdfplumber.open(filename) as pdf:
-#         page = pdf.pages[page_num]
-#         return page.extract_text()
-
-# def summarize_page(i, text):
-#     prompt = """Summarize the text in a detailed and concise manner."""
-#     # text = get_page_text(filename, page_num)
-#     client = OpenAI()
-#     # Check if text extraction was successful
-#     if text:
-#         response = client.chat.completions.create(
-#             model="gpt-4o",
-#             temperature=0.1,
-#             messages=[
-#                 {"role": "system", "content": prompt},
-#                 {"role": "user", "content": text}
-#             ]
-#         )
-#         return i, response.choices[0].message.content
-#     else:
-#         return i, "No text found on this page."
-
-# def summarize_large_text(texts):
-#     summary_parts = []
-#     num_pages = len(texts)
-#     with concurrent.futures.ThreadPoolExecutor() as executor:
-#         future_to_page = {executor.submit(summarize_page, i, texts[i]): i for i in range(num_pages)}
-#         for future in concurrent.futures.as_completed(future_to_page):
-#             i, page_summary = future.result()
-#             summary_parts.append((i, f"p{i+1}: {page_summary}"))
-
-#     # Sort the summary parts by page index and join them into a single string
-#     summary_parts.sort(key=lambda x: x[0])
-#     full_summary = "\n".join(part[1] for part in summary_parts)
-
-#     return full_summary
-
-# def summarize_pdf_parallel(filename):
-#     with pdfplumber.open(filename) as pdf:
-#         num_pages = len(pdf.pages)
-#         texts = [get_page_text(filename, i) for i in range(num_pages)]
-
-#     all_text = "\n".join(texts)
-
-#     if len(all_text) > 70000:
-#         return summarize_large_text(texts)
-#     else:
-#         return all_text
-
-# def summarize_file(filename):
-#     if filename.endswith(".pdf"):
-#         return summarize_pdf_parallel(filename)
-#     elif filename.endswith(".txt"):
-#         with open(filename, "r") as f:
-#             return f.read()
-#     elif filename.endswith(".docx"):
-#         return get_doc_text(filename)
-#     elif filename.endswith(".pptx"):
-#         return get_ppt_text(filename)
-#     else:
-#         raise ValueError("Unsupported file type")
 
 def summarize_chunk(chunk_text: str) -> str:
     llm = ChatOpenAI(temperature=0, model_name="gpt-4o")
@@ -331,7 +253,7 @@ def generate_audio(file=None, url=None, openai_api_key: str = None) -> bytes:
 
 
 demo = gr.Interface(
-    title="PDF to Podcast",
+    title="Antyhing to Podcast",
     description=Path("description.md").read_text(),
     fn=generate_audio,
     examples=[],
@@ -354,7 +276,7 @@ demo = gr.Interface(
     ],
     allow_flagging=False,
     clear_btn=None,
-    head=os.getenv("HEAD", "") + Path("head.html").read_text(),
+    head="Anythin Podcast",
     cache_examples="lazy",
     api_name=False,
 )

description.md

@@ -1,3 +1,3 @@
 <p style="text-align:center">
-  <strong>Convert any PDF into a podcast episode! Experience research papers, websites, and more in a whole new way.</strong>
+  <strong>Convert anything into a podcast episode! Experience research papers, websites, and more in a whole new way.</strong>
 </p>

head.html

@@ -1,49 +0,0 @@
-<!-- Primary Meta Tags -->
-<title>PDF to Podcast - Convert Your Documents to Audio</title>
-<meta name="title" content="PDF to Podcast - Convert Your Documents to Audio" />
-<meta
-  name="description"
-  content="Easily convert your PDF documents into audio podcasts. Perfect for listening on the go and making content more accessible."
-/>
-
-<!-- Open Graph / Facebook -->
-<meta property="og:type" content="website" />
-<meta property="og:url" content="https://pdf-to-podcast.com/" />
-<meta
-  property="og:title"
-  content="PDF to Podcast - Convert Your Documents to Audio"
-/>
-<meta
-  property="og:description"
-  content="Easily convert your PDF documents into audio podcasts. Perfect for listening on the go and making content more accessible."
-/>
-<meta
-  property="og:image"
-  content="https://pdf-to-podcast.com/static/logo.png"
-/>
-
-<!-- Twitter -->
-<meta property="twitter:card" content="summary_large_image" />
-<meta property="twitter:url" content="https://pdf-to-podcast.com/" />
-<meta
-  property="twitter:title"
-  content="PDF to Podcast - Convert Your Documents to Audio"
-/>
-<meta
-  property="twitter:description"
-  content="Easily convert your PDF documents into audio podcasts. Perfect for listening on the go and making content more accessible."
-/>
-<meta
-  property="twitter:image"
-  content="https://pdf-to-podcast.com/static/logo.png"
-/>
-
-<!-- Additional Meta Tags -->
-<meta name="viewport" content="width=device-width, initial-scale=1" />
-<meta charset="UTF-8" />
-<meta name="author" content="Stephan Fitzpatrick" />
-<meta
-  name="keywords"
-  content="PDF to Podcast, PDF to audio, document to podcast, audio conversion, podcast creation, accessible content"
-/>
-<link rel="icon" href="/static/icon.png" type="image/png" />

requirements.txt

@@ -4,16 +4,16 @@ pydantic
 google-generativeai
 loguru~=0.7
 pypdf~=4.1
-tenacity~=8.3
+tenacity==8.3.0
 sentry-sdk~=2.5
 granian~=1.4
 # PyPDF2
-pdfplumber
+pdfplumber==0.11.1
 openai
-# docx
+python-docx
 # python-pptx
-llmsherpa
-langchain_openai
-langchain_community
-langchain
-bs4
+llmsherpa~=0.1.4
+langchain_openai==0.1.8
+langchain_community==0.2.4
+langchain==0.2.5
+beautifulsoup4==4.12.3