Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -25,7 +25,6 @@ def process_pdf_to_epub(pdf_file, title, author):
|
|
25 |
doc = fitz.open(pdf_path)
|
26 |
num_pages = len(doc)
|
27 |
|
28 |
-
# Create EPUB book
|
29 |
book = epub.EpubBook()
|
30 |
book.set_identifier("id123456")
|
31 |
book.set_title(title)
|
@@ -38,12 +37,9 @@ def process_pdf_to_epub(pdf_file, title, author):
|
|
38 |
print(f"Processing page {page_num}...")
|
39 |
|
40 |
try:
|
41 |
-
# Render page to base64 image
|
42 |
image_base64 = render_pdf_to_base64png(pdf_path, page_num, target_longest_image_dim=1024)
|
43 |
anchor_text = get_anchor_text(pdf_path, page_num, pdf_engine="pdfreport", target_length=4000)
|
44 |
-
print(f"Anchor text for page {page_num}: {anchor_text}")
|
45 |
|
46 |
-
# New prompt format
|
47 |
prompt = (
|
48 |
"Below is the image of one page of a document, as well as some raw textual content that was previously "
|
49 |
"extracted for it. Just return the plain text representation of this document as if you were reading it naturally.\n"
|
@@ -99,28 +95,26 @@ def process_pdf_to_epub(pdf_file, title, author):
|
|
99 |
|
100 |
print(f"Decoded content for page {page_num}: {decoded}")
|
101 |
|
102 |
-
# Create chapter
|
103 |
chapter = epub.EpubHtml(title=f"Page {page_num}", file_name=f"page_{page_num}.xhtml", lang="en")
|
104 |
chapter.content = f"<h1>Page {page_num}</h1><p>{decoded}</p>"
|
105 |
book.add_item(chapter)
|
106 |
chapters.append(chapter)
|
107 |
|
108 |
-
# Save cover image from page 1
|
109 |
if page_num == 1:
|
110 |
cover_image = Image.open(BytesIO(base64.b64decode(image_base64)))
|
111 |
cover_io = BytesIO()
|
112 |
cover_image.save(cover_io, format='PNG')
|
113 |
book.set_cover("cover.png", cover_io.getvalue())
|
114 |
|
115 |
-
# Assemble EPUB
|
116 |
book.toc = tuple(chapters)
|
117 |
book.add_item(epub.EpubNcx())
|
118 |
book.add_item(epub.EpubNav())
|
119 |
book.spine = ['nav'] + chapters
|
120 |
|
121 |
-
|
122 |
-
|
123 |
-
|
|
|
124 |
|
125 |
# Gradio Interface
|
126 |
iface = gr.Interface(
|
|
|
25 |
doc = fitz.open(pdf_path)
|
26 |
num_pages = len(doc)
|
27 |
|
|
|
28 |
book = epub.EpubBook()
|
29 |
book.set_identifier("id123456")
|
30 |
book.set_title(title)
|
|
|
37 |
print(f"Processing page {page_num}...")
|
38 |
|
39 |
try:
|
|
|
40 |
image_base64 = render_pdf_to_base64png(pdf_path, page_num, target_longest_image_dim=1024)
|
41 |
anchor_text = get_anchor_text(pdf_path, page_num, pdf_engine="pdfreport", target_length=4000)
|
|
|
42 |
|
|
|
43 |
prompt = (
|
44 |
"Below is the image of one page of a document, as well as some raw textual content that was previously "
|
45 |
"extracted for it. Just return the plain text representation of this document as if you were reading it naturally.\n"
|
|
|
95 |
|
96 |
print(f"Decoded content for page {page_num}: {decoded}")
|
97 |
|
|
|
98 |
chapter = epub.EpubHtml(title=f"Page {page_num}", file_name=f"page_{page_num}.xhtml", lang="en")
|
99 |
chapter.content = f"<h1>Page {page_num}</h1><p>{decoded}</p>"
|
100 |
book.add_item(chapter)
|
101 |
chapters.append(chapter)
|
102 |
|
|
|
103 |
if page_num == 1:
|
104 |
cover_image = Image.open(BytesIO(base64.b64decode(image_base64)))
|
105 |
cover_io = BytesIO()
|
106 |
cover_image.save(cover_io, format='PNG')
|
107 |
book.set_cover("cover.png", cover_io.getvalue())
|
108 |
|
|
|
109 |
book.toc = tuple(chapters)
|
110 |
book.add_item(epub.EpubNcx())
|
111 |
book.add_item(epub.EpubNav())
|
112 |
book.spine = ['nav'] + chapters
|
113 |
|
114 |
+
# ✅ SAFELY write to a temp file in /tmp
|
115 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".epub", dir="/tmp") as tmp:
|
116 |
+
epub.write_epub(tmp.name, book)
|
117 |
+
return tmp.name
|
118 |
|
119 |
# Gradio Interface
|
120 |
iface = gr.Interface(
|