api: Remove dependency on exiftool
Use weasyprint to embed metadata
This commit is contained in:
TheOnlyWayUp
@@ -17,7 +17,6 @@ dependencies = [
|
||||
"aiohttp-client-cache[all]",
|
||||
"bs4>=0.0.2",
|
||||
"uvicorn>=0.32.1",
|
||||
"pyexiftool>=0.5.6",
|
||||
"weasyprint>=63.0",
|
||||
"jinja2>=3.1.6",
|
||||
]
|
||||
@@ -31,5 +30,6 @@ aiohttp-client-cache = { git = "https://github.com/TheOnlyWayUp/aiohttp-client-c
|
||||
[dependency-groups]
|
||||
dev = [
|
||||
"ipykernel>=6.29.5",
|
||||
"ipynb>=0.5.1",
|
||||
"ruff>=0.11.12",
|
||||
]
|
||||
|
||||
@@ -4,7 +4,6 @@ from pathlib import Path
|
||||
from tempfile import NamedTemporaryFile, _TemporaryFileWrapper
|
||||
|
||||
from bs4 import BeautifulSoup
|
||||
from exiftool import ExifTool
|
||||
from jinja2 import Template
|
||||
from weasyprint import CSS, HTML
|
||||
from weasyprint.text.fonts import FontConfiguration
|
||||
@@ -134,6 +133,12 @@ class PDFGenerator(AbstractGenerator):
|
||||
"book_title": self.story["title"],
|
||||
"cover": f"data:image/jpg;base64,{b64encode(self.cover).decode()}",
|
||||
"username": self.story["user"]["username"],
|
||||
"author_bio": self.story["user"]["description"],
|
||||
"tags": self.story["tags"],
|
||||
"created": self.story["createDate"],
|
||||
"modified": self.story["modifyDate"],
|
||||
"is_completed": self.story["completed"],
|
||||
"is_mature": self.story["mature"],
|
||||
"description": self.story["description"],
|
||||
"avatar": b64encode(self.author).decode(),
|
||||
"copyright": {
|
||||
@@ -160,44 +165,10 @@ class PDFGenerator(AbstractGenerator):
|
||||
self.book.name, stylesheets=[stylesheet_obj], font_config=font_config
|
||||
)
|
||||
|
||||
def add_metadata(self):
|
||||
"""Write metadata to generated PDF file at self.book, using ExifTool."""
|
||||
|
||||
clean_description = (
|
||||
self.story["description"].strip().replace("\n", "$/")
|
||||
) # exiftool doesn't parse \ns correctly, they support $/ for the same instead. `
` is another option.
|
||||
|
||||
metadata = {
|
||||
"Author": self.story["user"]["username"],
|
||||
"Title": self.story["title"],
|
||||
"Subject": clean_description,
|
||||
"CreationDate": self.story["createDate"],
|
||||
"ModDate": self.story["modifyDate"],
|
||||
"Keywords": ",".join(self.story["tags"]),
|
||||
"Language": self.story["language"]["name"],
|
||||
"Completed": self.story["completed"],
|
||||
"MatureContent": self.story["mature"],
|
||||
"Producer": "Dhanush Rambhatla (TheOnlyWayUp - https://rambhat.la) and WattpadDownloader",
|
||||
} # As per https://exiftool.org/TagNames/PDF.html
|
||||
|
||||
with ExifTool(config_file=DATA_PATH / "exiftool.config") as et:
|
||||
# Custom configuration adds Completed and MatureContent tags.
|
||||
# exiftool logger logs executed command
|
||||
et.execute(
|
||||
*(
|
||||
[f"-{key}={value}" for key, value in metadata.items()]
|
||||
+ [
|
||||
"-overwrite_original",
|
||||
self.book.file.name,
|
||||
]
|
||||
)
|
||||
)
|
||||
|
||||
def compile(self):
|
||||
parts = self.generate_chapters()
|
||||
self.populate_template(parts)
|
||||
self.generate_pdf()
|
||||
self.add_metadata()
|
||||
return True
|
||||
|
||||
def dump(self) -> BytesIO:
|
||||
|
||||
@@ -1,10 +1,26 @@
|
||||
<!DOCTYPE html>
|
||||
<html lang="{{ langcode }}">
|
||||
|
||||
|
||||
<head>
|
||||
|
||||
<title>{{ book_title }}</title>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
|
||||
<title>{{ book_title }}</title>
|
||||
<meta name=Subject content="{{description}}">
|
||||
<meta name=Author content="{{author}}">
|
||||
<meta name=Keywords content="{{tags}}">
|
||||
<meta name=Language content="{{langcode}}">
|
||||
<meta name=CreationDate content="{{created}}">
|
||||
<meta name=ModDate content="{{modified}}">
|
||||
<meta name=Generator content="Dhanush Rambhatla (TheOnlyWayUp - https://rambhat.la) and WattpadDownloader">
|
||||
|
||||
<meta name=completed content="{{is_completed}}">
|
||||
<meta name=maturecontent content="{{is_mature}}">
|
||||
|
||||
|
||||
</head>
|
||||
|
||||
<section class="fullpage">
|
||||
<img src="{{ cover }}" alt="Cover">
|
||||
@@ -23,10 +39,7 @@
|
||||
<div id="copyright-separator"></div>
|
||||
|
||||
{% if copyright.data %}
|
||||
<img src="data:image/jpg;base64,{{copyright.data}}"
|
||||
alt="{{copyright.name}}"
|
||||
width="88"
|
||||
height="31"
|
||||
<img src="data:image/jpg;base64,{{copyright.data}}" alt="{{copyright.name}}" width="88" height="31"
|
||||
id="copyright-license-image">
|
||||
{% endif %}
|
||||
|
||||
@@ -66,8 +79,9 @@ id="copyright-license-image">
|
||||
</h2>
|
||||
<hr id="author-divider">
|
||||
<p id="author-bio">
|
||||
{{ description }}
|
||||
{{ author_bio }}
|
||||
</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
</html>
|
||||
@@ -1,26 +0,0 @@
|
||||
|
||||
%Image::ExifTool::UserDefined = (
|
||||
'Image::ExifTool::XMP::xmp' => {
|
||||
Completed => {
|
||||
Writable => 'boolean', # Can be a boolean (True/False)
|
||||
Groups => { 2 => 'Content' },
|
||||
},
|
||||
MatureContent => {
|
||||
Writable => 'boolean', # Can be a boolean (True/False)
|
||||
Groups => { 2 => 'Content' },
|
||||
},
|
||||
},
|
||||
|
||||
'Image::ExifTool::IPTC::ApplicationRecord' => {
|
||||
161 => {
|
||||
Name => 'Completed',
|
||||
Format => 'string[0,16]', # Store as a string (e.g., "Yes"/"No")
|
||||
},
|
||||
162 => {
|
||||
Name => 'MatureContent',
|
||||
Format => 'string[0,16]', # Store as a string (e.g., "Yes"/"No")
|
||||
},
|
||||
},
|
||||
);
|
||||
|
||||
1; # End
|
||||
Generated
+459
-458
File diff suppressed because it is too large
Load Diff
Reference in New Issue
Block a user