orbitkit 0.8.39__tar.gz → 0.8.40__tar.gz

{orbitkit-0.8.39/orbitkit.egg-info → orbitkit-0.8.40}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: orbitkit
-Version: 0.8.39
+Version: 0.8.40
 Summary: This project is only for Orbit Tech internal use.
 Home-page: https://github.com/clown-0726/orbitkit
 Author: Lilu Cao

orbitkit-0.8.40/orbitkit/VERSION

@@ -0,0 +1 @@
+0.8.40

{orbitkit-0.8.39 → orbitkit-0.8.40}/orbitkit/airflow_handler/data_preprocessing.py

@@ -1,9 +1,13 @@
 import os
 import datetime
 from collections import defaultdict
+import asyncio
+from googletrans import Translator
 
 
 class DocumentProcessor:
+    _translator = Translator()
+
     AUDIO_SUFFIXES = [".mp3", ".wav", ".aac", ".wma", ".m4a"]
     VIDEO_SUFFIXES = [".mp4", ".avi", ".mkv", ".mov", ".wmv", ".flv", ".webm", ".mpeg", ".3gp", ".hevc"]
     PDF_SUFFIXES = [".pdf"]
@@ -17,6 +21,15 @@ class DocumentProcessor:
     def get_file_suffix(file_path):
         return f".{file_path.split('.')[-1]}".lower()
 
+    @classmethod
+    async def translate(cls, text, dest='en'):
+        result = await cls._translator.translate(text, dest=dest)
+        return result.text
+
+    @classmethod
+    def translate_sync(cls, text, dest='en'):
+        return asyncio.run(cls.translate(text, dest))
+
     @staticmethod
     def create_xbrl_template():
         return {
@@ -124,7 +137,8 @@ class DocumentProcessor:
                 _, ext = os.path.splitext(att['store_path'])
                 if ext in cls.AUDIO_SUFFIXES or ext in cls.VIDEO_SUFFIXES:
                     add_extends = {
-                        "title": doc['x_orbit_data']['report_title'],
+                        "original_title": doc['x_orbit_data']['report_title'],
+                        "title": cls.translate_sync(doc['x_orbit_data']['report_title']),
                         "published": doc['x_reported_at_utc_date'],
                         "tickers": [],
                         "perm_id_list": doc['x_orbit_data']['perm_id_list'],

{orbitkit-0.8.39 → orbitkit-0.8.40}/orbitkit/airflow_handler/file_handler_v2.py

@@ -5,7 +5,8 @@ from contextlib import contextmanager
 from typing import List, Dict, Tuple, Any, Optional, Union
 from sqlalchemy.dialects.postgresql import array
 from sqlalchemy.orm import sessionmaker, scoped_session
-from sqlalchemy import create_engine, MetaData, Table, select, Column, Integer, insert, and_, text, delete, update, desc
+from sqlalchemy import create_engine, MetaData, Table, select, Column, Integer, insert, and_, text, delete, update, \
+    desc, LABEL_STYLE_TABLENAME_PLUS_COL
 
 logger = logging.getLogger(__name__)
 
@@ -397,9 +398,9 @@ class FileFlowHandleV2:
             select(op_meta)
             .where(and_(*conditions)) if conditions else select(op_meta)
         )
-        meta_subquery = meta_subquery.order_by(desc(op_meta.c.priority)).limit(limit_size).subquery()
+        meta_subquery = meta_subquery.order_by(desc(op_meta.c.priority)).limit(limit_size).subquery(name="public_op_meta")
         join_stmt = meta_subquery.outerjoin(op_attachment, meta_subquery.c.id == op_attachment.c.meta_id)
-        stmt = select(meta_subquery, op_attachment).select_from(join_stmt)
+        stmt = (select(meta_subquery, op_attachment).select_from(join_stmt).set_label_style(LABEL_STYLE_TABLENAME_PLUS_COL))
 
         with self.session_scope() as session:
             result = session.execute(stmt).fetchall()
@@ -407,10 +408,10 @@ class FileFlowHandleV2:
             meta_map = {}
             for row in result:
                 row_dict = dict(row._mapping)
-                meta_id = row_dict['id']
+                meta_id = row_dict['public_op_meta_id']
 
-                meta_fields: Dict[str, Any] = {k: v for k, v in row_dict.items() if k in op_meta.c}
-                attachment_fields: Dict[str, Any] = {k: v for k, v in row_dict.items() if k in op_attachment.c}
+                meta_fields: Dict[str, Any] = {k.replace('public_op_meta_', ''): v for k, v in row_dict.items() if k.startswith('public_op_meta_')}
+                attachment_fields: Dict[str, Any] = {k.replace('public_op_attachment_', ''): v for k, v in row_dict.items() if k.startswith('public_op_attachment_')}
 
                 meta_fields['created_at'] = meta_fields['created_at'].isoformat()
                 meta_fields['updated_at'] = meta_fields['updated_at'].isoformat()

{orbitkit-0.8.39 → orbitkit-0.8.40/orbitkit.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: orbitkit
-Version: 0.8.39
+Version: 0.8.40
 Summary: This project is only for Orbit Tech internal use.
 Home-page: https://github.com/clown-0726/orbitkit
 Author: Lilu Cao

orbitkit-0.8.39/orbitkit/VERSION

@@ -1 +0,0 @@
-0.8.39