guillemhs/manga-translator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added pipeline

Browse Source
This commit is contained in:
Guillem Hernandez Sola
2026-04-16 19:58:05 +02:00
parent 5aa79d986a
commit 39765a6cf1
1 changed files with 46 additions and 86 deletions
Download Patch File Download Diff File Expand all files Collapse all files

126
pipeline.py
View File

@@ -4,67 +4,40 @@ pipeline.py
───────────────────────────────────────────────────────────────
Translation + render pipeline
Flow per page:
1) translate_manga_text() -> output.txt + bubbles.json (+ debug_clusters.png if DEBUG)
2) render_translations() -> page_translated.png
3) Pack CBZ with originals + rendered pages + text outputs
Folder structure:
<CHAPTER_DIR>/
├── 000.png
├── 001.png
└── translated/
├── 000/
│ ├── output.txt
│ ├── bubbles.json
│ ├── page_translated.png
│ └── debug_clusters.png (optional)
├── 001/
│ └── ...
└── ...
CBZ:
- pages/<original pages>
- rendered/<page_stem>_translated.png
- translations/<page_stem>_output.txt
Usage:
python pipeline.py /path/to/chapter/folder
"""
import os
import sys
import argparse
import zipfile
import importlib.util
from pathlib import Path
# ─────────────────────────────────────────────
# CONFIG
# ─────────────────────────────────────────────
CHAPTER_DIR = "/Users/guillemhernandezsola/Downloads/Dandadan_059_2022_Digital_1r0n"
OUTPUT_CBZ = "/Users/guillemhernandezsola/Downloads/Dandadan_059_2022_Digital_1r0n_translated.cbz"
SOURCE_LANG = "en"
TARGET_LANG = "ca"
# translator (NEW signature-compatible)
# Translator Settings
CONFIDENCE_THRESHOLD = 0.10
MIN_TEXT_LENGTH = 1
GAP_PX = "auto" # was cluster/proximity in old version
GAP_PX = "auto"
FILTER_SFX = True
QUALITY_THRESHOLD = 0.50
READING_MODE = "ltr"
DEBUG = True
# renderer
# Renderer Settings
RENDER_ENABLED = True
RENDER_OUTPUT_NAME = "page_translated.png"
# optional custom font list for renderer
FONT_CANDIDATES = [
FONT_CANDIDATES = [
"fonts/ComicNeue-Regular.ttf",
"fonts/ComicRelief-Regular.ttf"
]
# ─────────────────────────────────────────────
# DYNAMIC MODULE LOADER
# ─────────────────────────────────────────────
@@ -76,7 +49,6 @@ def load_module(name, filepath):
spec.loader.exec_module(module)
return module
# ─────────────────────────────────────────────
# HELPERS
# ─────────────────────────────────────────────
@@ -88,80 +60,61 @@ def sorted_pages(chapter_dir):
]
return sorted(pages, key=lambda p: p.stem)
def make_page_workdir(chapter_dir, page_stem):
workdir = Path(chapter_dir) / "translated" / page_stem
workdir.mkdir(parents=True, exist_ok=True)
return workdir
def pack_cbz(chapter_dir, translated_dir, output_cbz):
exts = {".jpg", ".jpeg", ".png", ".webp"}
pages = sorted(
[p for p in Path(chapter_dir).iterdir()
if p.is_file() and p.suffix.lower() in exts],
[p for p in Path(chapter_dir).iterdir() if p.is_file() and p.suffix.lower() in exts],
key=lambda p: p.stem
)
txts = sorted(
translated_dir.rglob("output.txt"),
key=lambda p: p.parent.name
)
rendered = sorted(
translated_dir.rglob(RENDER_OUTPUT_NAME),
key=lambda p: p.parent.name
)
txts = sorted(translated_dir.rglob("output.txt"), key=lambda p: p.parent.name)
rendered = sorted(translated_dir.rglob(RENDER_OUTPUT_NAME), key=lambda p: p.parent.name)
if not pages:
print("⚠️ No original pages found — CBZ not created.")
return
with zipfile.ZipFile(output_cbz, "w", compression=zipfile.ZIP_STORED) as zf:
# original pages
# Original pages
for img in pages:
arcname = f"pages/{img.name}"
zf.write(img, arcname)
print(f" 🖼 {arcname}")
# rendered pages
# Rendered pages
for rp in rendered:
arcname = f"rendered/{rp.parent.name}_translated.png"
zf.write(rp, arcname)
print(f" 🎨 {arcname}")
# text outputs
# Text outputs
for txt in txts:
arcname = f"translations/{txt.parent.name}_output.txt"
zf.write(txt, arcname)
print(f" 📄 {arcname}")
print(
f"\n✅ CBZ saved → {output_cbz} "
f"({len(pages)} original, {len(rendered)} rendered, {len(txts)} text)"
)
print(f"\n✅ CBZ saved → {output_cbz}")
print(f"📦 Contains: {len(pages)} original, {len(rendered)} rendered, {len(txts)} text files.")
# ─────────────────────────────────────────────
# PER-PAGE PIPELINE
# ─────────────────────────────────────────────
def process_page(page_path, workdir, translator_module, renderer_module):
"""
Runs translator + renderer for one page.
All generated files are written inside workdir.
"""
print(f"\n{'─' * 70}")
print(f"PAGE: {page_path.name}")
print(f"{'─' * 70}")
orig_dir = os.getcwd()
try:
# Isolate execution to the specific page's folder
os.chdir(workdir)
# 1) translate
# 1) Translate
translator_module.translate_manga_text(
image_path= str(page_path.resolve()),
image_path=str(page_path.resolve()),
source_lang=SOURCE_LANG,
target_lang=TARGET_LANG,
confidence_threshold=CONFIDENCE_THRESHOLD,
@@ -174,9 +127,9 @@ def process_page(page_path, workdir, translator_module, renderer_module):
reading_mode=READING_MODE,
debug=DEBUG
)
print(" ✅ translator done")
print(" ✅ Translator done")
# 2) render
# 2) Render
if RENDER_ENABLED:
renderer_module.render_translations(
input_image=str(page_path.resolve()),
@@ -185,7 +138,7 @@ def process_page(page_path, workdir, translator_module, renderer_module):
bubbles_file="bubbles.json",
font_candidates=FONT_CANDIDATES
)
print(" ✅ renderer done")
print(" ✅ Renderer done")
return True
@@ -196,44 +149,52 @@ def process_page(page_path, workdir, translator_module, renderer_module):
finally:
os.chdir(orig_dir)
# ─────────────────────────────────────────────
# MAIN
# ─────────────────────────────────────────────
def main():
parser = argparse.ArgumentParser(description="Manga Translation Pipeline")
parser.add_argument("chapter_dir", help="Path to the folder containing manga pages")
args = parser.parse_args()
chapter_dir = Path(args.chapter_dir).resolve()
output_cbz = chapter_dir.parent / f"{chapter_dir.name}_translated.cbz"
print("Loading modules...")
# Ensure we are loading from the directory where pipeline.py is located
script_dir = Path(__file__).parent
try:
translator = load_module("manga_translator", "manga-translator.py")
translator = load_module("manga_translator", str(script_dir / "manga-translator.py"))
except Exception as e:
print(f"❌ Could not load manga-translator.py: {e}")
sys.exit(1)
try:
renderer = load_module("manga_renderer", "manga-renderer.py")
renderer = load_module("manga_renderer", str(script_dir / "manga-renderer.py"))
except Exception as e:
print(f"❌ Could not load manga-renderer.py: {e}")
sys.exit(1)
pages = sorted_pages(CHAPTER_DIR)
pages = sorted_pages(chapter_dir)
if not pages:
print(f"❌ No images found in: {CHAPTER_DIR}")
print(f"❌ No images found in: {chapter_dir}")
sys.exit(1)
print(f"\n📖 Chapter : {CHAPTER_DIR}")
print(f"\n📖 Chapter : {chapter_dir}")
print(f" Pages : {len(pages)}")
print(f" Source : {SOURCE_LANG} → Target: {TARGET_LANG}")
print(f" Render : {'ON' if RENDER_ENABLED else 'OFF'}\n")
translated_dir = Path(CHAPTER_DIR) / "translated"
succeeded = []
failed = []
translated_dir = chapter_dir / "translated"
succeeded, failed = [], []
for i, page_path in enumerate(pages, start=1):
print(f"[{i}/{len(pages)}] {page_path.name}")
workdir = make_page_workdir(CHAPTER_DIR, page_path.stem)
ok = process_page(page_path, workdir, translator, renderer)
if ok:
print(f"[{i}/{len(pages)}] Processing...")
workdir = make_page_workdir(chapter_dir, page_path.stem)
if process_page(page_path, workdir, translator, renderer):
succeeded.append(page_path.name)
else:
failed.append(page_path.name)
@@ -248,8 +209,7 @@ def main():
print(f"{'═' * 70}\n")
print("Packing CBZ...")
pack_cbz(CHAPTER_DIR, translated_dir, OUTPUT_CBZ)
pack_cbz(chapter_dir, translated_dir, output_cbz)
if __name__ == "__main__":
main()