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

Added some fixes

Browse Source
This commit is contained in:
Guillem Hernandez Sola
2026-04-14 20:08:51 +02:00
parent 0069da706b
commit f95b7d32d4
5 changed files with 359 additions and 171 deletions
Download Patch File Download Diff File Expand all files Collapse all files

209
pipeline.py
View File

@@ -2,60 +2,76 @@
"""
pipeline.py
───────────────────────────────────────────────────────────────
Translation-only pipeline for Dandadan_059_2022_Digital
Translation + render pipeline
Flow per page:
1. Run translate_manga_text() → output.txt + bubbles.json
2. Copy original image to workdir for reference
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 produced:
Dandadan_059_2022_Digital_1r0n/
Folder structure:
<CHAPTER_DIR>/
├── 000.png
├── 001.png
└── translated/
├── 00/
│ ├── output.txt ← translations to review
│ ├── bubbles.json ← bubble boxes
│ └── debug_clusters.png ← cluster debug (if DEBUG=True)
├── 01/
├── 000/
│ ├── output.txt
│ ├── bubbles.json
│ ├── page_translated.png
│ └── debug_clusters.png (optional)
├── 001/
│ └── ...
└── ...
Dandadan_059_translated.cbz ← original pages + translations
zipped for reference
CBZ:
- pages/<original pages>
- rendered/<page_stem>_translated.png
- translations/<page_stem>_output.txt
"""
import os
import sys
import shutil
import zipfile
import importlib.util
from pathlib import Path
# ─────────────────────────────────────────────
# CONFIG — edit these as needed
# CONFIG
# ─────────────────────────────────────────────
CHAPTER_DIR = "/Users/guillemhernandezsola/Downloads/Spy_x_Family_076_2023_Digital_1r0n"
OUTPUT_CBZ = "/Users/guillemhernandezsola/Downloads/Spy_x_Family_076_2023_Digital_1r0n_translated.cbz"
SOURCE_LANG = "en"
TARGET_LANG = "ca"
CHAPTER_DIR = "/Users/guillemhernandezsola/Downloads/Dandadan_059_2022_Digital_1r0n"
OUTPUT_CBZ = "/Users/guillemhernandezsola/Downloads/Dandadan_059_2022_Digital_1r0n_translated.cbz"
# manga-translator.py settings
SOURCE_LANG = "en"
TARGET_LANG = "ca"
# translator (NEW signature-compatible)
CONFIDENCE_THRESHOLD = 0.10
MIN_TEXT_LENGTH = 2
CLUSTER_EPS = "auto"
PROXIMITY_PX = 80
MIN_TEXT_LENGTH = 1
GAP_PX = "auto" # was cluster/proximity in old version
FILTER_SFX = True
QUALITY_THRESHOLD = 0.5
UPSCALE_FACTOR = 2.5
BBOX_PADDING = 5
QUALITY_THRESHOLD = 0.50
READING_MODE = "ltr"
DEBUG = True
# renderer
RENDER_ENABLED = True
RENDER_OUTPUT_NAME = "page_translated.png"
# optional custom font list for renderer
FONT_CANDIDATES = [
"fonts/ComicNeue-Regular.ttf",
"fonts/ComicRelief-Regular.ttf"
]
# ─────────────────────────────────────────────
# DYNAMIC MODULE LOADER
# ─────────────────────────────────────────────
def load_module(name, filepath):
spec = importlib.util.spec_from_file_location(name, filepath)
spec = importlib.util.spec_from_file_location(name, filepath)
if spec is None or spec.loader is None:
raise FileNotFoundError(f"Cannot load spec for {filepath}")
module = importlib.util.module_from_spec(spec)
spec.loader.exec_module(module)
return module
@@ -65,10 +81,10 @@ def load_module(name, filepath):
# HELPERS
# ─────────────────────────────────────────────
def sorted_pages(chapter_dir):
exts = {".jpg", ".jpeg", ".png", ".webp"}
exts = {".jpg", ".jpeg", ".png", ".webp"}
pages = [
p for p in Path(chapter_dir).iterdir()
if p.suffix.lower() in exts
if p.is_file() and p.suffix.lower() in exts
]
return sorted(pages, key=lambda p: p.stem)
@@ -80,82 +96,97 @@ def make_page_workdir(chapter_dir, page_stem):
def pack_cbz(chapter_dir, translated_dir, output_cbz):
"""
Packs into CBZ:
- All original pages (from chapter_dir root)
- All output.txt (one per page subfolder)
Sorted by page stem for correct reading order.
"""
exts = {".jpg", ".jpeg", ".png", ".webp"}
pages = sorted(
exts = {".jpg", ".jpeg", ".png", ".webp"}
pages = sorted(
[p for p in Path(chapter_dir).iterdir()
if p.suffix.lower() in exts],
if p.is_file() and p.suffix.lower() in exts],
key=lambda p: p.stem
)
txts = sorted(
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
with zipfile.ZipFile(output_cbz, "w", compression=zipfile.ZIP_STORED) as zf:
# original pages
for img in pages:
arcname = f"pages/{img.name}"
zf.write(img, arcname)
print(f" 🖼 {arcname}")
# Translation text files
# rendered pages
for rp in rendered:
arcname = f"rendered/{rp.parent.name}_translated.png"
zf.write(rp, arcname)
print(f" 🎨 {arcname}")
# 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)} page(s), {len(txts)} translation(s))")
print(
f"\n✅ CBZ saved → {output_cbz} "
f"({len(pages)} original, {len(rendered)} rendered, {len(txts)} text)"
)
# ─────────────────────────────────────────────
# PER-PAGE PIPELINE
# ─────────────────────────────────────────────
def process_page(page_path, workdir, translator_module):
def process_page(page_path, workdir, translator_module, renderer_module):
"""
Runs translator for a single page.
All output files land in workdir.
Returns True on success, False on failure.
Runs translator + renderer for one page.
All generated files are written inside workdir.
"""
print(f"\n{'─'*60}")
print(f" PAGE: {page_path.name}")
print(f"{'─'*60}")
print(f"\n{'─' * 70}")
print(f"PAGE: {page_path.name}")
print(f"{'─' * 70}")
orig_dir = os.getcwd()
try:
# chdir into workdir so debug_clusters.png,
# temp files etc. all land there
os.chdir(workdir)
# 1) translate
translator_module.translate_manga_text(
image_path = str(page_path.resolve()),
source_lang = SOURCE_LANG,
target_lang = TARGET_LANG,
confidence_threshold = CONFIDENCE_THRESHOLD,
export_to_file = "output.txt",
export_bubbles_to = "bubbles.json",
min_text_length = MIN_TEXT_LENGTH,
cluster_eps = CLUSTER_EPS,
proximity_px = PROXIMITY_PX,
filter_sound_effects = FILTER_SFX,
quality_threshold = QUALITY_THRESHOLD,
upscale_factor = UPSCALE_FACTOR,
bbox_padding = BBOX_PADDING,
debug = DEBUG,
image_path= str(page_path.resolve()),
source_lang=SOURCE_LANG,
target_lang=TARGET_LANG,
confidence_threshold=CONFIDENCE_THRESHOLD,
min_text_length=MIN_TEXT_LENGTH,
gap_px=GAP_PX,
filter_sound_effects=FILTER_SFX,
quality_threshold=QUALITY_THRESHOLD,
export_to_file="output.txt",
export_bubbles_to="bubbles.json",
reading_mode=READING_MODE,
debug=DEBUG
)
print(" ✅ translator done")
# 2) render
if RENDER_ENABLED:
renderer_module.render_translations(
input_image=str(page_path.resolve()),
output_image=RENDER_OUTPUT_NAME,
translations_file="output.txt",
bubbles_file="bubbles.json",
font_candidates=FONT_CANDIDATES
)
print(" ✅ renderer done")
print(f" ✅ Translated → {workdir}")
return True
except Exception as e:
@@ -170,16 +201,20 @@ def process_page(page_path, workdir, translator_module):
# MAIN
# ─────────────────────────────────────────────
def main():
# ── Load translator module ────────────────────────────────────
print("Loading manga-translator.py...")
print("Loading modules...")
try:
translator = load_module(
"manga_translator", "manga-translator.py")
except FileNotFoundError as e:
print(f"❌ Could not load module: {e}")
translator = load_module("manga_translator", "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")
except Exception as e:
print(f"❌ Could not load manga-renderer.py: {e}")
sys.exit(1)
# ── Discover pages ────────────────────────────────────────────
pages = sorted_pages(CHAPTER_DIR)
if not pages:
print(f"❌ No images found in: {CHAPTER_DIR}")
@@ -187,33 +222,31 @@ def main():
print(f"\n📖 Chapter : {CHAPTER_DIR}")
print(f" Pages : {len(pages)}")
print(f" Source : {SOURCE_LANG} → Target: {TARGET_LANG}\n")
print(f" Source : {SOURCE_LANG} → Target: {TARGET_LANG}")
print(f" Render : {'ON' if RENDER_ENABLED else 'OFF'}\n")
# ── Process each page ─────────────────────────────────────────
translated_dir = Path(CHAPTER_DIR) / "translated"
succeeded = []
failed = []
failed = []
for i, page_path in enumerate(pages, start=1):
print(f"\n[{i}/{len(pages)}] {page_path.name}")
print(f"[{i}/{len(pages)}] {page_path.name}")
workdir = make_page_workdir(CHAPTER_DIR, page_path.stem)
ok = process_page(page_path, workdir, translator)
ok = process_page(page_path, workdir, translator, renderer)
if ok:
succeeded.append(page_path.name)
else:
failed.append(page_path.name)
# ── Summary ───────────────────────────────────────────────────
print(f"\n{'═'*60}")
print(f" PIPELINE COMPLETE")
print(f" ✅ {len(succeeded)} page(s) succeeded")
print(f"\n{'═' * 70}")
print("PIPELINE COMPLETE")
print(f"✅ {len(succeeded)} page(s) succeeded")
if failed:
print(f" ❌ {len(failed)} page(s) failed:")
print(f"❌ {len(failed)} page(s) failed:")
for f in failed:
print(f" • {f}")
print(f"{'═'*60}\n")
print(f" • {f}")
print(f"{'═' * 70}\n")
# ── Pack CBZ ──────────────────────────────────────────────────
print("Packing CBZ...")
pack_cbz(CHAPTER_DIR, translated_dir, OUTPUT_CBZ)