egg/Document_Translator
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

4th_fix time error

Browse Source
This commit is contained in:
beabigegg
2025-09-03 09:05:51 +08:00
parent e6e5332705
commit cce3fd4925
26 changed files with 2551 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
app/api/admin.py
View File

@@ -18,6 +18,7 @@ from app.utils.logger import get_logger
from app.models.user import User
from app.models.job import TranslationJob
from app.models.stats import APIUsageStats
from app.utils.timezone import format_taiwan_time
from app.models.log import SystemLog
from app.models.cache import TranslationCache
from sqlalchemy import func, desc
@@ -75,8 +76,8 @@ def get_system_stats():
'daily_stats': daily_stats,
'user_rankings': user_rankings_data,
'period': 'month',
'start_date': datetime.utcnow().isoformat(),
'end_date': datetime.utcnow().isoformat()
'start_date': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'end_date': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S")
}
))
@@ -359,7 +360,7 @@ def get_system_health():
try:
from datetime import datetime
status = {
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'status': 'healthy',
'services': {}
}
@@ -400,7 +401,7 @@ def get_system_health():
except Exception as e:
logger.error(f"Get system health error: {str(e)}")
return jsonify({
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'status': 'error',
'error': str(e)
}), 500
@@ -434,7 +435,7 @@ def get_system_metrics():
recent_counts = {status: count for status, count in recent_jobs}
metrics_data = {
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'jobs': {
'pending': job_counts.get('PENDING', 0),
'processing': job_counts.get('PROCESSING', 0),

9
app/api/health.py
View File

@@ -13,6 +13,7 @@ from flask import Blueprint, jsonify
from app.utils.helpers import create_response
from app.utils.logger import get_logger
from app.models.job import TranslationJob
from app.utils.timezone import format_taiwan_time, now_taiwan
health_bp = Blueprint('health', __name__, url_prefix='/health')
logger = get_logger(__name__)
@@ -23,7 +24,7 @@ def health_check():
"""系統健康檢查"""
try:
status = {
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'status': 'healthy',
'services': {}
}
@@ -108,7 +109,7 @@ def health_check():
except Exception as e:
logger.error(f"Health check error: {str(e)}")
return jsonify({
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'status': 'error',
'error': str(e)
}), 500
@@ -131,7 +132,7 @@ def get_metrics():
# 系統指標
metrics_data = {
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'jobs': {
'pending': job_counts.get('PENDING', 0),
'processing': job_counts.get('PROCESSING', 0),
@@ -217,6 +218,6 @@ def ping():
"""簡單的 ping 檢查"""
return jsonify({
'status': 'ok',
'timestamp': datetime.utcnow().isoformat(),
'timestamp': format_taiwan_time(datetime.utcnow(), "%Y-%m-%d %H:%M:%S"),
'message': 'pong'
})

2
app/config.py
View File

@@ -58,7 +58,7 @@ class Config:
CELERY_RESULT_SERIALIZER = 'json'
CELERY_ACCEPT_CONTENT = ['json']
CELERY_TIMEZONE = 'Asia/Taipei'
CELERY_ENABLE_UTC = True
CELERY_ENABLE_UTC = False # 改為 False，讓 Celery 使用本地時區
# LDAP 配置
LDAP_SERVER = os.environ.get('LDAP_SERVER')

11
app/models/job.py
View File

@@ -14,6 +14,7 @@ from datetime import datetime, timedelta
from sqlalchemy.sql import func
from sqlalchemy import event
from app import db
from app.utils.timezone import format_taiwan_time
class TranslationJob(db.Model):
@@ -80,10 +81,10 @@ class TranslationJob(db.Model):
'error_message': self.error_message,
'total_tokens': self.total_tokens,
'total_cost': float(self.total_cost) if self.total_cost else 0.0,
'processing_started_at': self.processing_started_at.isoformat() if self.processing_started_at else None,
'completed_at': self.completed_at.isoformat() if self.completed_at else None,
'created_at': self.created_at.isoformat() if self.created_at else None,
'updated_at': self.updated_at.isoformat() if self.updated_at else None
'processing_started_at': format_taiwan_time(self.processing_started_at, "%Y-%m-%d %H:%M:%S") if self.processing_started_at else None,
'completed_at': format_taiwan_time(self.completed_at, "%Y-%m-%d %H:%M:%S") if self.completed_at else None,
'created_at': format_taiwan_time(self.created_at, "%Y-%m-%d %H:%M:%S") if self.created_at else None,
'updated_at': format_taiwan_time(self.updated_at, "%Y-%m-%d %H:%M:%S") if self.updated_at else None
}
if include_files:
@@ -256,7 +257,7 @@ class JobFile(db.Model):
'filename': self.filename,
'file_path': self.file_path,
'file_size': self.file_size,
'created_at': self.created_at.isoformat() if self.created_at else None
'created_at': format_taiwan_time(self.created_at, "%Y-%m-%d %H:%M:%S") if self.created_at else None
}

3
app/models/stats.py
View File

@@ -11,6 +11,7 @@ Modified: 2024-01-28
from datetime import datetime, timedelta
from sqlalchemy.sql import func
from app import db
from app.utils.timezone import format_taiwan_time
class APIUsageStats(db.Model):
@@ -51,7 +52,7 @@ class APIUsageStats(db.Model):
'response_time_ms': self.response_time_ms,
'success': self.success,
'error_message': self.error_message,
'created_at': self.created_at.isoformat() if self.created_at else None
'created_at': format_taiwan_time(self.created_at, "%Y-%m-%d %H:%M:%S") if self.created_at else None
}
@classmethod

7
app/models/user.py
View File

@@ -11,6 +11,7 @@ Modified: 2024-01-28
from datetime import datetime, timedelta
from sqlalchemy.sql import func
from app import db
from app.utils.timezone import format_taiwan_time
class User(db.Model):
@@ -49,9 +50,9 @@ class User(db.Model):
'email': self.email,
'department': self.department,
'is_admin': self.is_admin,
'last_login': self.last_login.isoformat() if self.last_login else None,
'created_at': self.created_at.isoformat() if self.created_at else None,
'updated_at': self.updated_at.isoformat() if self.updated_at else None
'last_login': format_taiwan_time(self.last_login, "%Y-%m-%d %H:%M:%S") if self.last_login else None,
'created_at': format_taiwan_time(self.created_at, "%Y-%m-%d %H:%M:%S") if self.created_at else None,
'updated_at': format_taiwan_time(self.updated_at, "%Y-%m-%d %H:%M:%S") if self.updated_at else None
}
if include_stats:

93
app/services/document_processor.py
View File

@@ -577,56 +577,24 @@ def _insert_docx_translations(doc: docx.Document, segs: List[Segment],
continue
else:
# Normal paragraph (not in table cell) - enhanced logic from successful version
# Normal paragraph (not in table cell) - SIMPLIFIED FOR DEBUGGING
try:
# Check existing translations using the enhanced method
last = _find_last_inserted_after(p, limit=max(len(translations), 4))
# TEMPORARILY DISABLE existing translation check to force insertion
log(f"[DEBUG] 強制插入翻譯到段落: {seg.text[:30]}...")
# Check if all translations already exist
existing_texts = []
current_check = p
for _ in range(len(translations)):
try:
# Get the next sibling paragraph
next_sibling = current_check._element.getnext()
if next_sibling is not None and next_sibling.tag.endswith('}p'):
next_p = Paragraph(next_sibling, p._parent)
if _is_our_insert_block(next_p):
existing_texts.append(_p_text_with_breaks(next_p))
current_check = next_p
else:
break
else:
break
except Exception:
break
# Force all translations to be added
to_add = translations
# Skip if all translations already exist in order
if len(existing_texts) >= len(translations):
if all(_normalize_text(e) == _normalize_text(t) for e, t in zip(existing_texts[:len(translations)], translations)):
skip_cnt += 1
log(f"[SKIP] 段落已存在翻譯: {seg.text[:30]}...")
continue
# Determine which translations need to be added
to_add = []
for t in translations:
if not any(_normalize_text(t) == _normalize_text(e) for e in existing_texts):
to_add.append(t)
if not to_add:
skip_cnt += 1
log(f"[SKIP] 段落所有翻譯已存在: {seg.text[:30]}...")
continue
# Use enhanced insertion with proper positioning
anchor = last if last else p
# Use simple positioning - always insert after current paragraph
anchor = p
for block in to_add:
try:
log(f"[DEBUG] 嘗試插入: {block[:50]}...")
anchor = _append_after(anchor, block, italic=True, font_size_pt=INSERT_FONT_SIZE_PT)
log(f"[SUCCESS] _append_after成功插入")
except Exception as e:
log(f"[ERROR] 段落插入失敗: {e}, 嘗試簡化插入")
log(f"[ERROR] _append_after失敗: {e}, 嘗試簡化插入")
try:
# Fallback: simple append
if hasattr(p._parent, 'add_paragraph'):
@@ -640,7 +608,7 @@ def _insert_docx_translations(doc: docx.Document, segs: List[Segment],
continue
ok_cnt += 1
log(f"[SUCCESS] 段落插入 {len(to_add)} 個翻譯（交錯格式）")
log(f"[SUCCESS] 段落強制插入 {len(to_add)} 個翻譯")
except Exception as e:
log(f"[ERROR] 段落處理失敗: {e}, 跳過此段落")
@@ -686,6 +654,39 @@ class DocumentProcessor:
self.logger.error(f"Failed to extract DOCX segments from {file_path}: {str(e)}")
raise FileProcessingError(f"DOCX 文件分析失敗: {str(e)}")
def _rematch_segments_to_document(self, doc: docx.Document, old_segments: List[Segment]) -> List[Segment]:
"""Re-match segments from old document instance to new document instance."""
try:
# Extract fresh segments from the current document instance
fresh_segments = _collect_docx_segments(doc)
# Match old segments with fresh segments based on text content
matched_segments = []
for old_seg in old_segments:
# Find matching segment in fresh segments
matched = False
for fresh_seg in fresh_segments:
if (old_seg.kind == fresh_seg.kind and
old_seg.ctx == fresh_seg.ctx and
_normalize_text(old_seg.text) == _normalize_text(fresh_seg.text)):
matched_segments.append(fresh_seg)
matched = True
break
if not matched:
self.logger.warning(f"Failed to match segment: {old_seg.text[:50]}...")
# Still add the old segment but it might not work for insertion
matched_segments.append(old_seg)
self.logger.debug(f"Re-matched {len(matched_segments)} segments to current document")
return matched_segments
except Exception as e:
self.logger.error(f"Failed to re-match segments: {str(e)}")
# Return original segments as fallback
return old_segments
def insert_docx_translations(self, file_path: str, segments: List[Segment],
translation_map: Dict[Tuple[str, str], str],
target_languages: List[str], output_path: str) -> Tuple[int, int]:
@@ -693,11 +694,15 @@ class DocumentProcessor:
try:
doc = docx.Document(file_path)
# CRITICAL FIX: Re-match segments with the current document instance
# The original segments were extracted from a different document instance
matched_segments = self._rematch_segments_to_document(doc, segments)
def log_func(msg: str):
self.logger.debug(msg)
ok_count, skip_count = _insert_docx_translations(
doc, segments, translation_map, target_languages, log_func
doc, matched_segments, translation_map, target_languages, log_func
)
# Save the modified document

32
app/services/translation_service.py
View File

@@ -74,8 +74,11 @@ class DocxParser(DocumentParser):
def generate_translated_document(self, translations: Dict[str, List[str]],
target_language: str, output_dir: Path) -> str:
"""生成翻譯後的 DOCX 文件 - 使用增強的翻譯插入邏輯"""
"""生成翻譯後的 DOCX 文件 - 使用增強的翻譯插入邏輯（從快取讀取）"""
try:
from sqlalchemy import text as sql_text
from app import db
# 生成輸出檔名
output_filename = generate_filename(
self.file_path.name,
@@ -88,16 +91,29 @@ class DocxParser(DocumentParser):
# 提取段落資訊
segments = self.extract_segments_with_context()
# 建立翻譯映射
# 建立翻譯映射 - 從快取讀取而非使用傳入的translations參數
translation_map = {}
translated_texts = translations.get(target_language, [])
# 對應文字段落與翻譯
text_index = 0
logger.info(f"Building translation map for {len(segments)} segments in language {target_language}")
for seg in segments:
if text_index < len(translated_texts):
translation_map[(target_language, seg.text)] = translated_texts[text_index]
text_index += 1
# 從翻譯快取中查詢每個段落的翻譯
result = db.session.execute(sql_text("""
SELECT translated_text
FROM dt_translation_cache
WHERE source_text = :text AND target_language = :lang
ORDER BY created_at DESC
LIMIT 1
"""), {'text': seg.text, 'lang': target_language})
row = result.fetchone()
if row and row[0]:
translation_map[(target_language, seg.text)] = row[0]
logger.debug(f"Found translation for: {seg.text[:50]}...")
else:
logger.warning(f"No translation found for: {seg.text[:50]}...")
logger.info(f"Translation map built with {len(translation_map)} mappings")
# 使用增強的翻譯插入邏輯
ok_count, skip_count = self.processor.insert_docx_translations(