Merge branch 'develop'

fetching chat

MANIFEST.in

@@ -1 +1,3 @@
 include requirements.txt
+include requirements_test.txt
+

README.md

@@ -13,10 +13,14 @@ Other features:
 + Quick fetching of initial chat data by generating continuation params
 instead of web scraping.
 
+For more detailed information, see [wiki](https://github.com/taizan-hokuto/pytchat/wiki).
+
 ## Install
 ```python
 pip install pytchat
 ```
+## Demo
+![demo](https://taizan-hokuto.github.io/statics/demo.gif "demo")
 
 ## Examples
 ### on-demand mode
@@ -36,15 +40,15 @@ while chat.is_alive():
 from pytchat import LiveChat
 import time
 
-chat = LiveChat("G1w62uEMZ74", callback = func)
-while chat.is_alive():
-    #other background operation here.
-    time.sleep(3)
-
-def func(chatdata):
-    for c in chatdata.items:
+def main()
+    chat = LiveChat("G1w62uEMZ74", callback = func)
+    #other background operation.
+   
+#callback function is automatically called periodically.
+def func(data):
+    for c in data.items:
         print(f"{c.datetime} [{c.author.name}]-{c.message} {c.amountString}")
-        chat.tick()
+        data.tick()
 ```
 
 ### asyncio context:
@@ -54,21 +58,20 @@ import asyncio
 
 async def main():
     chat = LiveChatAsync("G1w62uEMZ74", callback = func)
-    while chat.is_alive():
-        #other background operation here.
-        await asyncio.sleep(3)
+    #other background operation.
 
-async def func(chat)
-    for c in chat.items:
+#callback function is automatically called periodically.
+async def func(data):
+    for c in data.items:
         print(f"{c.datetime} [{c.author.name}]-{c.message} {c.amountString}")
-        await chat.tick_async()
+        await data.tick_async()
 
 loop = asyncio.get_event_loop()
 loop.run_until_complete(main())
 ```
 
 
-### yt api compatible processor:
+### youtube api compatible processor:
 ```python
 from pytchat import LiveChat, CompatibleProcessor
 
@@ -85,10 +88,32 @@ while chat.is_alive():
             time.sleep(polling/len(data["items"]))
 
 ```
+### replay:
+```python
+from pytchat import ReplayChatAsync
+import asyncio
 
+async def main():
+    chat = ReplayChatAsync("G1w62uEMZ74", seektime = 1000, callback = func)
+        #other background operation here.
 
-## Chatdata Structure of Default Processor
-Structure of each item which got from items() function.
+#callback function is automatically called periodically.
+async def func(data):
+    for count in range(0,len(data.items)):
+        c= data.items[count]
+        if count!=len(data.items):
+            tick=data.items[count+1].timestamp -data.items[count].timestamp
+        else:
+            tick=0
+        print(f"<{c.elapsedTime}> [{c.author.name}]-{c.message} {c.amountString}")
+        await asyncio.sleep(tick/1000)
+
+loop = asyncio.get_event_loop()
+loop.run_until_complete(main())
+```
+
+## Structure of Default Processor
+Each item can be got with items() function.
 <table>
   <tr>
     <th>name</th>
@@ -110,6 +135,11 @@ Structure of each item which got from items() function.
     <td>str</td>
     <td>emojis are represented by ":(shortcut text):"</td>
   </tr>
+  <tr>
+    <td>messageEx</td>
+    <td>str</td>
+    <td>list of message texts and emoji URLs.</td>
+  </tr>
   <tr>
     <td>timestamp</td>
     <td>int</td>
@@ -118,7 +148,11 @@ Structure of each item which got from items() function.
   <tr>
     <td>datetime</td>
     <td>str</td>
-    <td></td>
+    <td>ex. "2019-10-10 12:34:56"</td>
+  </tr>
+    <td>elapsedTime</td>
+    <td>str</td>
+    <td>elapsed time. (ex. "1:02:27") *Replay Only.</td>
   </tr>
   <tr>
     <td>amountValue</td>
@@ -133,7 +167,12 @@ Structure of each item which got from items() function.
   <tr>
     <td>currency</td>
     <td>str</td>
-    <td>ex. "USD"</td>
+    <td><a href="https://en.wikipedia.org/wiki/ISO_4217">ISO 4217 currency codes</a> (ex. "USD")</td>
+  </tr>
+  <tr>
+    <td>bgColor</td>
+    <td>int</td>
+    <td>RGB Int</td>
   </tr>
   <tr>
     <td>author</td>
@@ -157,7 +196,7 @@ Structure of author object.
   <tr>
     <td>channelId</td>
     <td>str</td>
-    <td></td>
+    <td>*chatter's channel ID. NOT broadcasting video's channel ID.</td>
   </tr>
   <tr>
     <td>channelUrl</td>
@@ -166,12 +205,12 @@ Structure of author object.
   </tr>
   <tr>
     <td>imageUrl</td>
-    <td></td>
+    <td>str</td>
     <td></td>
   </tr>
   <tr>
     <td>badgeUrl</td>
-    <td></td>
+    <td>str</td>
     <td></td>
   </tr>
   <tr>

pytchat/__init__.py

@@ -1,18 +1,21 @@
 """
-pytchat is a python library for fetching youtube live chat.
+pytchat is a python library for fetching youtube live chat without using yt api, Selenium, or BeautifulSoup.
 """
 __copyright__    = 'Copyright (C) 2019 taizan-hokuto'
-__version__      = '0.0.1.7'
+__version__      = '0.0.3.4'
 __license__      = 'MIT'
 __author__       = 'taizan-hokuto'
 __author_email__ = '55448286+taizan-hokuto@users.noreply.github.com'
-__url__          = 'https://github.com/taizan-hokuto'
+__url__          = 'https://github.com/taizan-hokuto/pytchat'
 
 __all__ = ["core_async","core_multithread","processors"]
 
 from .api import (
     LiveChat,
     LiveChatAsync,
+    ReplayChat,
+    ReplayChatAsync,
+    ChatProcessor,
     CompatibleProcessor, 
     SimpleDisplayProcessor,
     JsonfileArchiveProcessor

pytchat/api.py

@@ -1,5 +1,8 @@
-from .core_async.livechat import LiveChatAsync
 from .core_multithread.livechat import LiveChat
+from .core_async.livechat import LiveChatAsync
+from .core_multithread.replaychat import ReplayChat
+from .core_async.replaychat import ReplayChatAsync
+from .processors.chat_processor import ChatProcessor
 from .processors.default.processor import DefaultProcessor
 from .processors.compatible.processor import CompatibleProcessor
 from .processors.simple_display_processor import SimpleDisplayProcessor

pytchat/core_async/buffer.py

@@ -2,7 +2,7 @@
 import asyncio
 class Buffer(asyncio.Queue):
     '''
-    チャットデータを格納するバッファの役割を持つLIFOキュー
+    チャットデータを格納するバッファの役割を持つFIFOキュー
 
     Parameter
     ---------

pytchat/core_async/listen_manager.py

@@ -1,184 +0,0 @@
-import asyncio
-from .listener import AsyncListener
-from .. import config
-from .. import mylogger
-import datetime
-import os
-import aiohttp
-import signal
-import threading
-from .buffer import Buffer
-from  concurrent.futures import CancelledError
-logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
-
-class ListenManager:
-
-    '''
-    動画IDまたは動画IDのリストを受け取り、
-    動画IDに対応したListenerを生成・保持する。
-
-    #Attributes
-    ----------
-    _listeners: dict  
-        ListenManegerがつかんでいるListener達のリスト.  
-        key:動画ID value:動画IDに対応するListener
-    _queue: Queue
-        動画IDを外部から受け渡しするためのキュー
-        _queueが空である間は、ノンブロッキングで他のタスクを実行
-        _queueに動画IDが投入されると、_dequeueメソッドで
-        直ちにListenerを生成し返す。
-    _event: threading.Event
-        キーボードのCTRL+Cを検知するためのEventオブジェクト
-    '''
-    def __init__(self,interruptable = True):
-        #チャット監視中の動画リスト
-        self._listeners={}
-        self._tasks = []
-        #外部からvideoを受け取るためのキュー
-        self._queue = asyncio.Queue()
-        self._event = threading.Event()
-        self._ready_queue()
-        self._is_alive = True
-        #キーボードのCtrl+cを押したとき、_hundler関数を呼び出すように設定
-        signal.signal(signal.SIGINT, (lambda a, b: self._handler(self._event, a, b)))
-
-    def is_alive(self)->bool:
-        '''
-        ListenManagerが稼働中であるか。
-            True->稼働中
-            False->Ctrl+Cが押されて終了
-        '''
-        logger.debug(f'check is_alive() :{self._is_alive}')
-        return self._is_alive
-    
-    def _handler(self, event, sig, handler):
-        '''
-        Ctrl+Cが押下されたとき、終了フラグをセットする。
-        '''
-        logger.debug('Ctrl+c pushed')
-        self._is_alive = False
-        logger.debug('terminating listeners.')
-        for listener in self._listeners.values():
-            listener.terminate()
-        logger.debug('end.')
-
-
-    def _ready_queue(self):
-        #loop = asyncio.get_event_loop()
-        self._tasks.append(
-            asyncio.create_task(self._dequeue())
-            )
-
-
-    async def set_video_ids(self,video_ids:list):
-        for video_id in video_ids:
-            if video_id:
-                await self._queue.put(video_id)
-
-
-    async def get_listener(self,video_id) -> AsyncListener:
-        return await self._create_listener(video_id)
-            
-    # async def getlivechat(self,video_id):
-    #     '''
-    #     指定された動画IDのチャットデータを返す
-
-    #     Parameter
-    #     ----------
-    #     video_id: str
-    #         動画ID
-
-    #     Return
-    #     ----------
-    #         引数で受け取った動画IDに対応する
-    #         Listenerオブジェクトへの参照
-            
-    #     '''
-    #     logger.debug('manager get/create listener')
-    #     listener = await self._create_listener(video_id)
-    #     '''
-    #     上が完了しないうちに、下が呼び出される
-    #     '''
-    #     if not listener._initialized:
-    #         await asyncio.sleep(2)
-    #         return []
-    #     if listener:
-    #         #listener._isfirstrun=False
-    #         return await listener.getlivechat()
-        
-
-
-    async def _dequeue(self):
-        '''
-        キューに入った動画IDを
-        Listener登録に回す。
-        
-        '''
-        while True:
-            video_id = await self._queue.get()
-            #listenerを登録、タスクとして実行する
-            logger.debug(f'deque got [{video_id}]')
-            await self._create_listener(video_id)
-
-    async def _create_listener(self, video_id) -> AsyncListener:
-        '''
-        Listenerを作成しチャット取得中リストに加え、
-        Listenerを返す
-        '''
-        if video_id is None or not isinstance(video_id, str):
-            raise TypeError('video_idは文字列でなければなりません')
-        if video_id in self._listeners:
-            return self._listeners[video_id]
-        else:
-            #listenerを登録する
-            listener = AsyncListener(video_id,interruptable = False,buffer = Buffer())
-            self._listeners.setdefault(video_id,listener)
-            #task = asyncio.ensure_future(listener.initialize())
-            #await asyncio.gather(listener.initialize())
-            #task.add_done_callback(self.finish)
-            #await listener.initialize()
-            #self._tasks.append(task)
-
-            return listener
-
- 
-    def finish(self,sender):
-        try:
-            if sender.result():
-                video_id = sender.result()[0]
-                message = sender.result()[1]
-            
-                #listener終了時のコールバック        
-                #sender.result()[]でデータを取得できる
-                logger.info(f'終了しました VIDEO_ID:[{video_id}] message:{message}')
-                #logger.info(f'終了しました')
-                if video_id in self._listeners:
-                    self._listeners.pop(video_id)
-        except CancelledError:
-            logger.debug('cancelled.')
-
-    def get_listeners(self):
-        return self._listeners
-
-    def shutdown(self):
-        '''
-        ListenManegerを終了する
-        '''
-        logger.debug("start shutdown")
-        self._is_alive =False
-        try:
-            #Listenerを停止する。
-            for listener in self._listeners.values():
-                listener.terminate()
-            #taskをキャンセルする。
-            for task in self._tasks:
-                if not task.done():
-                    #print(task)
-                    task.cancel()
-        except Exception as er:
-            logger.info(str(er),type(er))
-
-        logger.debug("finished.")
-
-    def get_tasks(self):
-        return self._tasks

pytchat/core_async/livechat.py

@@ -1,21 +1,19 @@
-import aiohttp, asyncio, async_timeout
+import aiohttp, asyncio
 import datetime
 import json
 import random
 import signal
-import threading
 import time
 import traceback
 import urllib.parse
 from aiohttp.client_exceptions import ClientConnectorError
 from concurrent.futures import CancelledError
 from .buffer import Buffer
-from .parser import Parser
+from ..parser.live import Parser
 from .. import config
 from .. import mylogger
 from ..exceptions  import ChatParseException,IllegalFunctionCall
 from ..paramgen    import liveparam
-
 from ..processors.default.processor import DefaultProcessor
 
 logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
@@ -81,7 +79,7 @@ class LiveChatAsync:
         self._exception_handler = exception_handler
         self._direct_mode = direct_mode
         self._is_alive   = True
-
+        self._parser = Parser()
         self._setup()
         
         if not LiveChatAsync._setup_finished:
@@ -124,7 +122,7 @@ class LiveChatAsync:
 
     async def _startlisten(self):
         """最初のcontinuationパラメータを取得し、
-        _listenループを開始する
+        _listenループのタスクを作成し開始する
         """
         initial_continuation = await self._get_initial_continuation()
         if initial_continuation is None:
@@ -164,7 +162,7 @@ class LiveChatAsync:
                     livechat_json = (await 
                       self._get_livechat_json(continuation, session, headers)
                     )
-                    metadata, chatdata =  Parser.parse( livechat_json )
+                    metadata, chatdata =  self._parser.parse( livechat_json )
                     timeout = metadata['timeoutMs']/1000
                     chat_component = {
                         "video_id" : self.video_id,
@@ -287,13 +285,4 @@ class LiveChatAsync:
         logger.debug(f"残っているタスクを終了しています")
         await asyncio.gather(*tasks,return_exceptions=True)
         loop = asyncio.get_event_loop()
-        loop.stop()
-
-
-
-  
-    
-
-
-
-
+        loop.stop()

pytchat/core_async/replaychat.py

@@ -0,0 +1,311 @@
+import aiohttp, asyncio
+import datetime
+import json
+import random
+import signal
+import time
+import traceback
+import urllib.parse
+from aiohttp.client_exceptions import ClientConnectorError
+from concurrent.futures import CancelledError
+from queue import Queue
+from .buffer import Buffer
+from ..parser.replay import Parser
+from .. import config
+from .. import mylogger
+from ..exceptions  import ChatParseException,IllegalFunctionCall
+from ..paramgen    import arcparam
+from ..processors.default.processor import DefaultProcessor
+
+logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
+MAX_RETRY = 10
+headers = config.headers
+
+
+
+class ReplayChatAsync:
+    '''asyncio(aiohttp)を利用してYouTubeのチャットデータを取得する。
+
+    Parameter
+    ---------
+    video_id : str
+        動画ID
+
+    seektime : int
+        リプレイするチャットデータの開始時間（秒）
+
+    processor : ChatProcessor
+        チャットデータを加工するオブジェクト
+
+    buffer : Buffer(maxsize:20[default])
+        チャットデータchat_componentを格納するバッファ。
+        maxsize : 格納できるchat_componentの個数
+        default値20個。1個で約5~10秒分。
+
+    interruptable : bool
+        Ctrl+Cによる処理中断を行うかどうか。
+
+    callback : func
+        _listen()関数から一定間隔で自動的に呼びだす関数。
+
+    done_callback : func
+        listener終了時に呼び出すコールバック。
+
+    exception_handler : func
+        例外を処理する関数
+
+    direct_mode : bool
+        Trueの場合、bufferを使わずにcallbackを呼ぶ。
+        Trueの場合、callbackの設定が必須
+        (設定していない場合IllegalFunctionCall例外を発生させる）  
+
+    Attributes
+    ---------
+    _is_alive : bool
+        チャット取得を停止するためのフラグ
+    '''
+
+    _setup_finished = False
+
+    def __init__(self, video_id,
+                seektime = 0,
+                processor = DefaultProcessor(),
+                buffer = None,
+                interruptable = True,
+                callback = None,
+                done_callback = None,
+                exception_handler = None,
+                direct_mode = False):
+        self.video_id  = video_id
+        self.seektime = seektime
+        self.processor = processor
+        self._buffer = buffer
+        self._callback = callback
+        self._done_callback = done_callback
+        self._exception_handler = exception_handler
+        self._direct_mode = direct_mode
+        self._is_alive   = True
+        self._parser = Parser()
+        self._pauser = Queue()
+        self._pauser.put_nowait(None)
+        self._setup()
+
+        if not ReplayChatAsync._setup_finished:
+            ReplayChatAsync._setup_finished = True
+            if exception_handler == None:
+                self._set_exception_handler(self._handle_exception)
+            else:
+                self._set_exception_handler(exception_handler)
+            if interruptable:
+                signal.signal(signal.SIGINT,
+                  (lambda a, b:asyncio.create_task(  
+                  ReplayChatAsync.shutdown(None,signal.SIGINT,b))
+                ))
+
+    def _setup(self):
+        #direct modeがTrueでcallback未設定の場合例外発生。
+        if self._direct_mode:
+            if self._callback is None:
+                raise IllegalFunctionCall(
+                    "direct_mode=Trueの場合callbackの設定が必須です。")
+        else:
+            #direct modeがFalseでbufferが未設定ならばデフォルトのbufferを作成
+            if self._buffer is None:
+                self._buffer = Buffer(maxsize = 20)
+                #callbackが指定されている場合はcallbackを呼ぶループタスクを作成
+            if self._callback is None:
+                pass 
+            else:
+                #callbackを呼ぶループタスクの開始
+                loop = asyncio.get_event_loop()
+                loop.create_task(self._callback_loop(self._callback))
+        #_listenループタスクの開始
+        loop = asyncio.get_event_loop()
+        listen_task = loop.create_task(self._startlisten())
+        #add_done_callbackの登録
+        if self._done_callback is None:
+            listen_task.add_done_callback(self.finish)
+        else:
+            listen_task.add_done_callback(self._done_callback)
+
+    async def _startlisten(self):
+        """最初のcontinuationパラメータを取得し、
+        _listenループのタスクを作成し開始する
+        """
+        initial_continuation = await self._get_initial_continuation()
+        if initial_continuation is None:
+            self.terminate()
+            logger.debug(f"[{self.video_id}]No initial continuation.")
+            return
+        await self._listen(initial_continuation)
+   
+    async def _get_initial_continuation(self):
+        ''' チャットデータ取得に必要な最初のcontinuationを取得する。'''
+        try:    
+            initial_continuation = arcparam.get(self.video_id,self.seektime)
+        except ChatParseException as e:
+            self.terminate()
+            logger.debug(f"[{self.video_id}]Error:{str(e)}")
+            return
+        except KeyError:
+            logger.debug(f"[{self.video_id}]KeyError:"
+                         f"{traceback.format_exc(limit = -1)}")
+            self.terminate()
+            return
+        return initial_continuation
+
+    async def _listen(self, continuation):
+        ''' continuationに紐付いたチャットデータを取得し
+        Bufferにチャットデータを格納、
+        次のcontinuaitonを取得してループする。
+
+        Parameter
+        ---------
+        continuation : str
+            次のチャットデータ取得に必要なパラメータ
+        '''
+        try:
+            async with aiohttp.ClientSession() as session:
+                while(continuation and self._is_alive):
+                    if self._pauser.empty():
+                        #pause
+                        await self._pauser.get()
+                        #resume
+                        #prohibit from blocking by putting None into _pauser.
+                        self._pauser.put_nowait(None)
+                    livechat_json = (await
+                      self._get_livechat_json(continuation, session, headers)
+                    )
+                    metadata, chatdata =  self._parser.parse( livechat_json )
+                    timeout = metadata['timeoutMs']/1000
+                    chat_component = {
+                        "video_id" : self.video_id,
+                        "timeout"  : timeout,
+                        "chatdata" : chatdata
+                    }
+                    time_mark =time.time()
+                    if self._direct_mode:
+                        await self._callback(
+                            self.processor.process([chat_component])
+                            )
+                    else:
+                        await self._buffer.put(chat_component)
+                    diff_time = timeout - (time.time()-time_mark)
+                    await asyncio.sleep(diff_time)       
+                    continuation = metadata.get('continuation')  
+        except ChatParseException as e:
+            logger.info(f"{str(e)}（video_id:\"{self.video_id}\"）")
+            return            
+        except (TypeError , json.JSONDecodeError) :
+            logger.error(f"{traceback.format_exc(limit = -1)}")
+            return
+        
+        logger.debug(f"[{self.video_id}]チャット取得を終了しました。")
+
+    async def _get_livechat_json(self, continuation, session, headers):
+        '''
+        チャットデータが格納されたjsonデータを取得する。
+        '''
+        continuation = urllib.parse.quote(continuation)
+        livechat_json = None
+        status_code = 0
+        url =(
+            f"https://www.youtube.com/live_chat_replay/get_live_chat_replay?"
+            f"continuation={continuation}&pbj=1")
+        for _ in range(MAX_RETRY + 1):
+            async with session.get(url ,headers = headers) as resp:
+                try:
+                    text = await resp.text()
+                    status_code = resp.status
+                    livechat_json = json.loads(text)
+                    break
+                except (ClientConnectorError,json.JSONDecodeError) :
+                    await asyncio.sleep(1)
+                    continue
+        else:
+            logger.error(f"[{self.video_id}]"
+                    f"Exceeded retry count. status_code={status_code}")
+            return None
+        return livechat_json
+  
+    async def _callback_loop(self,callback):
+        """ コンストラクタでcallbackを指定している場合、バックグラウンドで
+        callbackに指定された関数に一定間隔でチャットデータを投げる。        
+        
+        Parameter
+        ---------
+        callback : func
+            加工済みのチャットデータを渡す先の関数。
+        """
+        while self.is_alive():
+            items = await self._buffer.get()
+            data = self.processor.process(items)
+            await callback(data)
+
+    async def get(self):
+        """ bufferからデータを取り出し、processorに投げ、
+        加工済みのチャットデータを返す。
+        
+        Returns
+             : Processorによって加工されたチャットデータ
+        """
+        if self._callback is None:
+            items = await self._buffer.get()
+            return  self.processor.process(items)
+        raise IllegalFunctionCall(
+            "既にcallbackを登録済みのため、get()は実行できません。")
+
+    def pause(self):
+        if not self._pauser.empty():
+            self._pauser.get()
+
+    def resume(self):
+        if self._pauser.empty():
+            self._pauser.put_nowait(None)
+        
+
+    def is_alive(self):
+        return self._is_alive
+
+    def finish(self,sender):
+        '''Listener終了時のコールバック'''
+        try: 
+            self.terminate()
+        except CancelledError:
+            logger.debug(f'[{self.video_id}]cancelled:{sender}')
+
+    def terminate(self):
+        '''
+        Listenerを終了する。
+        '''
+        self._is_alive = False
+        if self._direct_mode == False:
+            #bufferにダミーオブジェクトを入れてis_alive()を判定させる
+            self._buffer.put_nowait({'chatdata':'','timeout':1}) 
+        logger.info(f'終了しました:[{self.video_id}]')
+  
+    @classmethod
+    def _set_exception_handler(cls, handler):
+        loop = asyncio.get_event_loop()
+        #default handler: cls._handle_exception
+        loop.set_exception_handler(handler)
+    
+    @classmethod
+    def _handle_exception(cls, loop, context):
+        #msg = context.get("exception", context["message"])
+        if not isinstance(context["exception"],CancelledError):
+            logger.error(f"Caught exception: {context}")
+        loop= asyncio.get_event_loop()
+        loop.create_task(cls.shutdown(None,None,None))
+
+    @classmethod
+    async def shutdown(cls, event, sig = None, handler=None):
+        logger.debug("シャットダウンしています")
+        tasks = [t for t in asyncio.all_tasks() if t is not
+        asyncio.current_task()]
+        [task.cancel() for task in tasks]
+
+        logger.debug(f"残っているタスクを終了しています")
+        await asyncio.gather(*tasks,return_exceptions=True)
+        loop = asyncio.get_event_loop()
+        loop.stop()

pytchat/core_multithread/buffer.py

@@ -3,7 +3,7 @@ import queue
 
 class Buffer(queue.Queue):
     '''
-    チャットデータを格納するバッファの役割を持つLIFOキュー
+    チャットデータを格納するバッファの役割を持つFIFOキュー
 
     Parameter
     ---------

pytchat/core_multithread/livechat.py

@@ -3,13 +3,12 @@ import datetime
 import json
 import random
 import signal
-import threading
 import time
 import traceback
 import urllib.parse
 from concurrent.futures import CancelledError, ThreadPoolExecutor
 from .buffer import Buffer
-from .parser import Parser
+from ..parser.live import Parser
 from .. import config
 from .. import mylogger
 from ..exceptions  import ChatParseException,IllegalFunctionCall
@@ -58,7 +57,7 @@ class LiveChat:
         チャットデータ取得ループ（_listen）用のスレッド
 
     _is_alive : bool
-        チャット取得を終了したか
+        チャット取得を停止するためのフラグ
     '''
 
     _setup_finished = False
@@ -143,7 +142,7 @@ class LiveChat:
 
     def _listen(self, continuation):
         ''' continuationに紐付いたチャットデータを取得し
-        にチャットデータを格納、
+        BUfferにチャットデータを格納、
         次のcontinuaitonを取得してループする
 
         Parameter
@@ -158,7 +157,6 @@ class LiveChat:
                       self._get_livechat_json(continuation, session, headers)
                     )
                     metadata, chatdata =  self._parser.parse( livechat_json )
-                    #チャットデータを含むコンポーネントを組み立ててbufferに投入する
                     timeout = metadata['timeoutMs']/1000
                     chat_component = {
                         "video_id" : self.video_id,
@@ -172,16 +170,12 @@ class LiveChat:
                             )
                     else:
                         self._buffer.put(chat_component)
-                    #次のchatを取得するまでsleepする
                     diff_time = timeout - (time.time()-time_mark)
                     if diff_time < 0 : diff_time=0
                     time.sleep(diff_time)        
-                    #次のチャットデータのcontinuationパラメータを取り出す。
                     continuation = metadata.get('continuation')  
-    
-                    #whileループ先頭に戻る
         except ChatParseException as e:
-            logger.error(f"{str(e)}（動画ID:\"{self.video_id}\"）")
+            logger.info(f"{str(e)}（video_id:\"{self.video_id}\"）")
             return            
         except (TypeError , json.JSONDecodeError) :
             logger.error(f"{traceback.format_exc(limit = -1)}")

pytchat/core_multithread/replaychat.py

@@ -0,0 +1,286 @@
+import requests
+import datetime
+import json
+import random
+import signal
+import time
+import traceback
+import urllib.parse
+from concurrent.futures import CancelledError, ThreadPoolExecutor
+from queue import Queue
+from .buffer import Buffer
+from ..parser.replay import Parser
+from .. import config
+from .. import mylogger
+from ..exceptions  import ChatParseException,IllegalFunctionCall
+from ..paramgen    import arcparam
+from ..processors.default.processor import DefaultProcessor
+
+logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
+MAX_RETRY = 10
+headers = config.headers
+
+
+
+class ReplayChat:
+    ''' スレッドプールを利用してYouTubeのライブ配信のチャットデータを取得する
+
+    Parameter
+    ---------
+    video_id : str
+        動画ID
+
+    seektime : int
+        リプレイするチャットデータの開始時間（秒）
+
+    processor : ChatProcessor
+        チャットデータを加工するオブジェクト
+
+    buffer : Buffer(maxsize:20[default])
+        チャットデータchat_componentを格納するバッファ。
+        maxsize : 格納できるchat_componentの個数
+        default値20個。1個で約5~10秒分。
+
+    interruptable : bool
+        Ctrl+Cによる処理中断を行うかどうか。
+
+    callback : func
+        _listen()関数から一定間隔で自動的に呼びだす関数。
+
+    done_callback : func
+        listener終了時に呼び出すコールバック。
+
+    direct_mode : bool
+        Trueの場合、bufferを使わずにcallbackを呼ぶ。
+        Trueの場合、callbackの設定が必須
+        (設定していない場合IllegalFunctionCall例外を発生させる）  
+
+    Attributes
+    ---------
+    _executor : ThreadPoolExecutor
+        チャットデータ取得ループ（_listen）用のスレッド
+
+    _is_alive : bool
+        チャット取得を終了したか
+    '''
+
+    _setup_finished = False
+    #チャット監視中のListenerのリスト
+    _listeners= []
+    def __init__(self, video_id,
+                seektime = 0,
+                processor = DefaultProcessor(),
+                buffer = Buffer(maxsize = 20),
+                interruptable = True,
+                callback = None,
+                done_callback = None,
+                direct_mode = False
+                ):
+        self.video_id  = video_id
+        self.seektime = seektime
+        self.processor = processor
+        self._buffer = buffer
+        self._callback = callback
+        self._done_callback = done_callback
+        self._executor = ThreadPoolExecutor(max_workers=2)
+        self._direct_mode = direct_mode
+        self._is_alive   = True
+        self._parser = Parser()
+        self._pauser = Queue()
+        self._pauser.put_nowait(None)
+       
+        self._setup()
+        if not ReplayChat._setup_finished:
+            ReplayChat._setup_finished = True
+            if interruptable:
+                signal.signal(signal.SIGINT,  (lambda a, b:  
+                (ReplayChat.shutdown(None,signal.SIGINT,b))
+                ))
+        ReplayChat._listeners.append(self)
+
+    def _setup(self):
+        #direct modeがTrueでcallback未設定の場合例外発生。
+        if self._direct_mode:
+            if self._callback is None:
+                raise IllegalFunctionCall(
+                    "direct_mode=Trueの場合callbackの設定が必須です。")
+        else:
+            #direct modeがFalseでbufferが未設定ならばデフォルトのbufferを作成
+            if self._buffer is None:
+                self._buffer = Buffer(maxsize = 20)
+                #callbackが指定されている場合はcallbackを呼ぶループタスクを作成
+            if self._callback is None:
+                pass 
+            else:
+                #callbackを呼ぶループタスクの開始
+                self._executor.submit(self._callback_loop,self._callback)
+        #_listenループタスクの開始
+        listen_task = self._executor.submit(self._startlisten)
+        #add_done_callbackの登録
+        if self._done_callback is None:
+            listen_task.add_done_callback(self.finish)
+        else:
+            listen_task.add_done_callback(self._done_callback)
+
+    def _startlisten(self):
+        """最初のcontinuationパラメータを取得し、
+        _listenループのタスクを作成し開始する
+        """
+        initial_continuation = self._get_initial_continuation()
+        if initial_continuation is None:
+            self.terminate()
+            logger.debug(f"[{self.video_id}]No initial continuation.")
+            return
+        self._listen(initial_continuation)
+   
+    def _get_initial_continuation(self):
+        ''' チャットデータ取得に必要な最初のcontinuationを取得する。'''
+        try:    
+            initial_continuation = arcparam.get(self.video_id,self.seektime)
+        except ChatParseException as e:
+            self.terminate()
+            logger.debug(f"[{self.video_id}]Error:{str(e)}")
+            return
+        except KeyError:
+            logger.debug(f"[{self.video_id}]KeyError:"
+                         f"{traceback.format_exc(limit = -1)}")
+            self.terminate()
+            return
+        return initial_continuation
+
+    def _listen(self, continuation):
+        ''' continuationに紐付いたチャットデータを取得し
+        にチャットデータを格納、
+        次のcontinuaitonを取得してループする
+
+        Parameter
+        ---------
+        continuation : str
+            次のチャットデータ取得に必要なパラメータ
+        '''
+        try:
+            with requests.Session() as session:
+                while(continuation and self._is_alive):
+                    if self._pauser.empty():
+                        #pause
+                        self._pauser.get()
+                        #resume
+                        #prohibit from blocking by putting None into _pauser.
+                        self._pauser.put_nowait(None)
+                    livechat_json = (
+                      self._get_livechat_json(continuation, session, headers)
+                    )
+                    metadata, chatdata =  self._parser.parse( livechat_json )
+                    timeout = metadata['timeoutMs']/1000
+                    chat_component = {
+                        "video_id" : self.video_id,
+                        "timeout"  : timeout,
+                        "chatdata" : chatdata
+                    }
+                    time_mark =time.time()
+                    if self._direct_mode:
+                        self._callback(
+                            self.processor.process([chat_component])
+                            )
+                    else:
+                        self._buffer.put(chat_component)
+                    diff_time = timeout - (time.time()-time_mark)
+                    if diff_time < 0 : diff_time=0
+                    time.sleep(diff_time)        
+                    continuation = metadata.get('continuation')  
+        except ChatParseException as e:
+            logger.error(f"{str(e)}（動画ID:\"{self.video_id}\"）")
+            return            
+        except (TypeError , json.JSONDecodeError) :
+            logger.error(f"{traceback.format_exc(limit = -1)}")
+            return
+        
+        logger.debug(f"[{self.video_id}]チャット取得を終了しました。")
+
+    def _get_livechat_json(self, continuation, session, headers):
+        '''
+        チャットデータが格納されたjsonデータを取得する。
+        '''
+        continuation = urllib.parse.quote(continuation)
+        livechat_json = None
+        status_code = 0
+        url =(
+            f"https://www.youtube.com/live_chat_replay/get_live_chat_replay?"
+            f"continuation={continuation}&pbj=1")
+        for _ in range(MAX_RETRY + 1):
+            with session.get(url ,headers = headers) as resp:
+                try:
+                    text = resp.text
+                    status_code = resp.status_code
+                    livechat_json = json.loads(text)
+                    break
+                except json.JSONDecodeError :
+                    time.sleep(1)
+                    continue
+        else:
+            logger.error(f"[{self.video_id}]"
+                    f"Exceeded retry count. status_code={status_code}")
+            return None
+        return livechat_json
+  
+    def _callback_loop(self,callback):
+        """ コンストラクタでcallbackを指定している場合、バックグラウンドで
+        callbackに指定された関数に一定間隔でチャットデータを投げる。        
+        
+        Parameter
+        ---------
+        callback : func
+            加工済みのチャットデータを渡す先の関数。
+        """
+        while self.is_alive():
+            items = self._buffer.get()
+            data = self.processor.process(items)
+            callback(data)
+
+    def get(self):
+        """ bufferからデータを取り出し、processorに投げ、
+        加工済みのチャットデータを返す。
+        
+        Returns
+             : Processorによって加工されたチャットデータ
+        """
+        if self._callback is None:
+            items = self._buffer.get()
+            return  self.processor.process(items)
+        raise IllegalFunctionCall(
+            "既にcallbackを登録済みのため、get()は実行できません。")
+
+    def pause(self):
+        if not self._pauser.empty():
+            self._pauser.get()
+
+    def resume(self):
+        if self._pauser.empty():
+            self._pauser.put_nowait(None)
+        
+
+    def is_alive(self):
+        return self._is_alive
+
+    def finish(self,sender):
+        '''Listener終了時のコールバック'''
+        try: 
+            self.terminate()
+        except CancelledError:
+            logger.debug(f'[{self.video_id}]cancelled:{sender}')
+
+    def terminate(self):
+        '''
+        Listenerを終了する。
+        '''
+        self._is_alive = False
+        if self._direct_mode == False:
+            #bufferにダミーオブジェクトを入れてis_alive()を判定させる
+            self._buffer.put({'chatdata':'','timeout':1}) 
+        logger.info(f'終了しました:[{self.video_id}]')
+  
+    @classmethod
+    def shutdown(cls, event, sig = None, handler=None):
+        logger.debug("シャットダウンしています")
+        for t in ReplayChat._listeners:
+            t._is_alive = False

pytchat/mylogger.py

@@ -17,7 +17,7 @@ def get_logger(modname,mode=logging.DEBUG):
     logger.addHandler(handler1)
     #create handler2 for recording log file
     if mode <= logging.DEBUG:
-        handler2 = logging.FileHandler(filename="log.txt")
+        handler2 = logging.FileHandler(filename="log.txt", encoding='utf-8')
         handler2.setLevel(logging.ERROR)
         handler2.setFormatter(my_formatter)
 

pytchat/paramgen/arcparam.py

@@ -0,0 +1,120 @@
+from base64 import urlsafe_b64encode as b64enc
+from functools import reduce
+import calendar, datetime, pytz
+import math
+import random
+import urllib.parse
+
+'''
+Generate continuation parameter of youtube replay chat.
+
+Author: taizan-hokuto (2019) @taizan205
+
+ver 0.0.1 2019.10.05
+'''
+
+def _gen_vid(video_id):
+    """generate video_id parameter.
+    Parameter
+    ---------
+        video_id : str
+
+    Return
+    ---------
+    byte[] : base64 encoded video_id parameter.
+    """
+    header_magic = b'\x0A\x0F\x1A\x0D\x0A'
+    header_id =  video_id.encode()
+    header_sep_1 = b'\x1A\x13\xEA\xA8\xDD\xB9\x01\x0D\x0A\x0B'
+    header_terminator = b'\x20\x01'
+
+    item = [
+        header_magic,
+        _nval(len(header_id)),
+        header_id,
+        header_sep_1,
+        header_id,
+        header_terminator
+    ]
+
+    return urllib.parse.quote(
+        b64enc(reduce(lambda x, y: x+y, item)).decode()
+    ).encode()
+
+def _nval(val):
+    """convert value to byte array"""
+    if val<0: raise ValueError
+    buf = b''
+    while val >> 7:
+        m = val & 0xFF | 0x80
+        buf += m.to_bytes(1,'big')
+        val >>= 7
+    buf += val.to_bytes(1,'big')
+    return buf
+
+
+def _tzparity(video_id,times):
+    t=0
+    for i,s in enumerate(video_id):
+        ss = ord(s)
+        if(ss % 2 == 0):
+            t += ss*(12-i)
+        else:
+            t ^= ss*i
+
+    return ((times^t) % 2).to_bytes(1,'big')
+
+
+def get(video_id, seektime = 0, topchatonly = False):
+    switch_01 = b'\x04' if topchatonly else b'\x01'
+
+
+    if seektime < 0:
+        raise ValueError('seektime is 0 or positive number.')
+    if seektime == 0:  
+        times =_nval(1)
+        switch = b'\x04'       
+    else:
+        times =_nval(int(seektime*1000000))
+        switch = b'\x03'
+    parity = _tzparity(video_id, seektime)
+
+    header_magic= b'\xA2\x9D\xB0\xD3\x04'
+    sep_0       = b'\x1A'
+    vid         = _gen_vid(video_id)
+    time_tag    = b'\x28'
+    timestamp1  = times
+    sep_1       = b'\x30\x00\x38\x00\x40\x00\x48'
+    sep_2       = b'\x52\x1C\x08\x00\x10\x00\x18\x00\x20\x00'
+    chkstr      = b'\x2A\x0E\x73\x74\x61\x74\x69\x63\x63\x68\x65\x63\x6B\x73\x75\x6D\x40'
+    sep_3       = b'\x00\x58\x03\x60'
+    sep_4       = b'\x68'+parity+b'\x72\x04\x08'
+    sep_5       = b'\x10'+parity+b'\x78\x00'
+    body = [
+        sep_0,
+        _nval(len(vid)),
+        vid,
+        time_tag,
+        timestamp1,
+        sep_1,
+        switch,
+        sep_2,
+        chkstr,
+        sep_3,
+        switch_01,
+        sep_4,
+        switch_01,
+        sep_5
+    ]
+
+    body = reduce(lambda x, y: x+y, body)
+    
+    return  urllib.parse.quote(
+                b64enc( header_magic +
+                        _nval(len(body)) +
+                        body
+                ).decode()
+            )
+
+
+

pytchat/paramgen/liveparam.py

@@ -1,11 +1,16 @@
 from base64 import urlsafe_b64encode as b64enc
 from functools import reduce
 import calendar, datetime, pytz
-import math
 import random
 import urllib.parse
 
+'''
+Generate continuation parameter of youtube live chat.
 
+Author: taizan-hokuto (2019) @taizan205
+
+ver 0.0.1 2019.10.05
+'''
 def _gen_vid(video_id):
     """generate video_id parameter.
     Parameter
@@ -39,6 +44,17 @@ def _gen_vid(video_id):
         b64enc(reduce(lambda x, y: x+y, item)).decode()
     ).encode()
 
+def _tzparity(video_id,times):
+    t=0
+    for i,s in enumerate(video_id):
+        ss = ord(s)
+        if(ss % 2 == 0):
+            t += ss*(12-i)
+        else:
+            t ^= ss*i
+
+    return ((times^t) % 2).to_bytes(1,'big')
+
 def _nval(val):
     """convert value to byte array"""
     if val<0: raise ValueError
@@ -53,15 +69,16 @@ def _nval(val):
 def _build(video_id, _ts1, _ts2, _ts3, _ts4, _ts5, topchatonly = False):
     #_short_type2
     switch_01 = b'\x04' if topchatonly else b'\x01'
-    header_magic= b'\xD2\x87\xCC\xC8\x03'
+    parity = _tzparity(video_id, _ts1^_ts2^_ts3^_ts4^_ts5)
 
+    header_magic= b'\xD2\x87\xCC\xC8\x03'
     sep_0       = b'\x1A'
     vid         = _gen_vid(video_id)
     time_tag    = b'\x28'
     timestamp1  = _nval(_ts1)
     sep_1       = b'\x30\x00\x38\x00\x40\x02\x4A'
     un_len      = b'\x2B'
-    sep_2       = b'\x08\x00\x10\x00\x18\x00\x20\x00'
+    sep_2       = b'\x08'+parity+b'\x10\x00\x18\x00\x20\x00'
     chkstr      = b'\x2A\x0E\x73\x74\x61\x74\x69\x63\x63\x68\x65\x63\x6B\x73\x75\x6D'
     sep_3       = b'\x3A\x00\x40\x00\x4A'
     sep_4_len   = b'\x02'

pytchat/parser/replay.py

@@ -11,8 +11,7 @@ logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
 
 
 class Parser:
-    @classmethod
-    def parse(cls, jsn):
+    def parse(self, jsn):
         if jsn is None: 
             return {'timeoutMs':0,'continuation':None},[]
         if jsn['response']['responseContext'].get('errors'):
@@ -26,15 +25,26 @@ class Parser:
         cont = contents['liveChatContinuation']['continuations'][0]
         if cont is None:
             raise NoContinuationsException('Continuationがありません。')
-        metadata = (cont.get('invalidationContinuationData')  or
-                    cont.get('timedContinuationData')         or
-                    cont.get('reloadContinuationData')
-                    )
+        metadata = cont.get('liveChatReplayContinuationData')
         if metadata is None:
             unknown = list(cont.keys())[0]
             if unknown:
-                logger.error(f"Received unknown continuation type:{unknown}")
                 metadata = cont.get(unknown)
-        metadata.setdefault('timeoutMs', 10000)
-        chatdata = contents['liveChatContinuation'].get('actions')
+        
+        actions = contents['liveChatContinuation'].get('actions')
+        if actions is None:
+            raise NoContentsException('チャットデータを取得できませんでした。')
+        interval = self.get_interval(actions)
+        metadata.setdefault("timeoutMs",interval)
+        chatdata = [action["replayChatItemAction"]["actions"][0] for action in actions]
         return metadata, chatdata
+
+    def get_interval(self, actions: list):
+        if actions is None:
+            return 0
+        start = int(actions[0]["replayChatItemAction"]["videoOffsetTimeMsec"])
+        last = int(actions[-1]["replayChatItemAction"]["videoOffsetTimeMsec"])
+        return (last - start)
+
+
+

pytchat/processors/chat_processor.py

@@ -5,11 +5,12 @@ class ChatProcessor:
     '''
     def process(self, chat_components: list):
         '''
-        チャットデータの加工を表すインターフェース
-        Listenerから呼び出される。        
+        チャットデータの加工を表すインターフェース。
+        LiveChatオブジェクトから呼び出される。 
+               
         Parameter
         ----------
-        chat_components: list<component>
+        chat_components: [LIST:component]
             component : dict {
                 "video_id" : str
                     動画ID

pytchat/processors/compatible/parser.py

@@ -1,43 +0,0 @@
-
-from .renderer.textmessage import LiveChatTextMessageRenderer
-from .renderer.paidmessage import LiveChatPaidMessageRenderer
-from .renderer.paidsticker import LiveChatPaidStickerRenderer
-from .renderer.legacypaid import LiveChatLegacyPaidMessageRenderer
-
-def parse(sitem):
-
-    action = sitem.get("addChatItemAction")
-    if action:
-        item = action.get("item")
-    if item is None: return None
-    rd={}
-    try:
-        renderer = get_renderer(item)
-        if renderer == None:
-            return None
-
-        rd["kind"] = "youtube#liveChatMessage"
-        rd["etag"] = ""
-        rd["id"] = 'LCC.' + renderer.get_id()
-        rd["snippet"]       = renderer.get_snippet()
-        rd["authorDetails"] = renderer.get_authordetails()
-    except (KeyError,TypeError,AttributeError) as e:
-        print(f"------{str(type(e))}-{str(e)}----------")
-        print(sitem)
-        return None
-    
-    return rd        
-
-def get_renderer(item):
-    if item.get("liveChatTextMessageRenderer"):
-        renderer = LiveChatTextMessageRenderer(item)
-    elif item.get("liveChatPaidMessageRenderer"):
-        renderer = LiveChatPaidMessageRenderer(item)
-    elif item.get( "liveChatPaidStickerRenderer"):
-        renderer = LiveChatPaidStickerRenderer(item)
-    elif item.get("liveChatLegacyPaidMessageRenderer"):
-        renderer = LiveChatLegacyPaidMessageRenderer(item)
-    else:
-        renderer = None
-    return renderer
-

pytchat/processors/compatible/processor.py

@@ -1,11 +1,15 @@
-from . import parser
-import json
-import os
-import traceback
 import datetime
 import time
-class CompatibleProcessor():
-    
+from .renderer.textmessage import LiveChatTextMessageRenderer
+from .renderer.paidmessage import LiveChatPaidMessageRenderer
+from .renderer.paidsticker import LiveChatPaidStickerRenderer
+from .renderer.legacypaid import LiveChatLegacyPaidMessageRenderer
+from ... import mylogger
+from ... import config
+logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
+
+class CompatibleProcessor:
+
     def process(self, chat_components: list):
 
         chatlist = []
@@ -26,7 +30,7 @@ class CompatibleProcessor():
                     if action.get('addChatItemAction') is None: continue
                     if action['addChatItemAction'].get('item') is None: continue
 
-                    chat = parser.parse(action)
+                    chat = self.parse(action)
                     if chat:
                         chatlist.append(chat)
         ret["pollingIntervalMillis"] = int(timeout*1000)
@@ -36,4 +40,42 @@ class CompatibleProcessor():
         }
         ret["items"] = chatlist
 
-        return ret
+        return ret
+
+    def parse(self, sitem):
+
+        action = sitem.get("addChatItemAction")
+        if action:
+            item = action.get("item")
+        if item is None: return None
+        rd={}
+        try:
+            renderer = self.get_renderer(item)
+            if renderer == None:
+                return None
+
+            rd["kind"] = "youtube#liveChatMessage"
+            rd["etag"] = ""
+            rd["id"] = 'LCC.' + renderer.get_id()
+            rd["snippet"]       = renderer.get_snippet()
+            rd["authorDetails"] = renderer.get_authordetails()
+        except (KeyError,TypeError,AttributeError) as e:
+            logger.error(f"Error: {str(type(e))}-{str(e)}")
+            logger.error(f"item: {sitem}")
+            return None
+        
+        return rd        
+
+    def get_renderer(self, item):
+        if item.get("liveChatTextMessageRenderer"):
+            renderer = LiveChatTextMessageRenderer(item)
+        elif item.get("liveChatPaidMessageRenderer"):
+            renderer = LiveChatPaidMessageRenderer(item)
+        elif item.get( "liveChatPaidStickerRenderer"):
+            renderer = LiveChatPaidStickerRenderer(item)
+        elif item.get("liveChatLegacyPaidMessageRenderer"):
+            renderer = LiveChatLegacyPaidMessageRenderer(item)
+        else:
+            renderer = None
+        return renderer
+

pytchat/processors/default/parser.py

@@ -1,39 +0,0 @@
-
-from .renderer.textmessage import LiveChatTextMessageRenderer
-from .renderer.paidmessage import LiveChatPaidMessageRenderer
-from .renderer.paidsticker import LiveChatPaidStickerRenderer
-from .renderer.legacypaid import LiveChatLegacyPaidMessageRenderer
-
-def parse(sitem):
-
-    action = sitem.get("addChatItemAction")
-    if action:
-        item = action.get("item")
-    if item is None: return None
-    try:
-        renderer = get_renderer(item)
-        if renderer == None:
-            return None
-
-        renderer.get_snippet()
-        renderer.get_authordetails()
-    except (KeyError,TypeError,AttributeError) as e:
-        print(f"------{str(type(e))}-{str(e)}----------")
-        print(sitem)
-        return None
-    
-    return renderer        
-
-def get_renderer(item):
-    if item.get("liveChatTextMessageRenderer"):
-        renderer = LiveChatTextMessageRenderer(item)
-    elif item.get("liveChatPaidMessageRenderer"):
-        renderer = LiveChatPaidMessageRenderer(item)
-    elif item.get( "liveChatPaidStickerRenderer"):
-        renderer = LiveChatPaidStickerRenderer(item)
-    elif item.get("liveChatLegacyPaidMessageRenderer"):
-        renderer = LiveChatLegacyPaidMessageRenderer(item)
-    else:
-        renderer = None
-    return renderer
-

pytchat/processors/default/processor.py

@@ -1,7 +1,12 @@
-from . import parser
 import asyncio
 import time
-
+from .renderer.textmessage import LiveChatTextMessageRenderer
+from .renderer.paidmessage import LiveChatPaidMessageRenderer
+from .renderer.paidsticker import LiveChatPaidStickerRenderer
+from .renderer.legacypaid import LiveChatLegacyPaidMessageRenderer
+from ... import config
+from ... import mylogger
+logger = mylogger.get_logger(__name__,mode=config.LOGGER_MODE)
 
 class Chatdata:
     def __init__(self,chatlist:list, timeout:float):
@@ -10,13 +15,13 @@ class Chatdata:
     
     def tick(self):
         if self.interval == 0:
-            time.sleep(3)
+            time.sleep(1)
             return
         time.sleep(self.interval/len(self.items))
 
     async def tick_async(self):
         if self.interval == 0:
-            await asyncio.sleep(0.5)
+            await asyncio.sleep(1)
             return
         await asyncio.sleep(self.interval/len(self.items))
 
@@ -37,8 +42,39 @@ class DefaultProcessor:
                     if action.get('addChatItemAction') is None: continue
                     if action['addChatItemAction'].get('item') is None: continue
 
-                    chat = parser.parse(action)
+                    chat = self._parse(action)
                     if chat:
                         chatlist.append(chat)
         return Chatdata(chatlist, float(timeout))
   
+
+    def _parse(self, sitem):
+
+        action = sitem.get("addChatItemAction")
+        if action:
+            item = action.get("item")
+        if item is None: return None
+        try:
+            renderer = self._get_renderer(item)
+            if renderer == None:
+                return None
+
+            renderer.get_snippet()
+            renderer.get_authordetails()
+        except (KeyError,TypeError,AttributeError) as e:
+            logger.error(f"{str(type(e))}-{str(e)} sitem:{str(sitem)}")
+            return None
+        return renderer        
+
+    def _get_renderer(self, item):
+        if item.get("liveChatTextMessageRenderer"):
+            renderer = LiveChatTextMessageRenderer(item)
+        elif item.get("liveChatPaidMessageRenderer"):
+            renderer = LiveChatPaidMessageRenderer(item)
+        elif item.get( "liveChatPaidStickerRenderer"):
+            renderer = LiveChatPaidStickerRenderer(item)
+        elif item.get("liveChatLegacyPaidMessageRenderer"):
+            renderer = LiveChatLegacyPaidMessageRenderer(item)
+        else:
+            renderer = None
+        return renderer

pytchat/processors/default/renderer/base.py

@@ -13,8 +13,14 @@ class BaseRenderer:
         self.id = self.renderer.get('id')
         timestampUsec = int(self.renderer.get("timestampUsec",0))
         self.timestamp = int(timestampUsec/1000)
+        tst = self.renderer.get("timestampText")
+        if tst:
+            self.elapsedTime = tst.get("simpleText")
+        else:
+            self.elapsedTime = ""
         self.datetime = self.get_datetime(timestampUsec)
         self.message = self.get_message(self.renderer)
+        self.messageEx = self.get_message_ex(self.renderer)
         self.id =  self.renderer.get('id')
         self.amountValue= 0.0
         self.amountString = ""
@@ -49,6 +55,19 @@ class BaseRenderer:
                             message += r.get('text','')
         return message
 
+    def get_message_ex(self,renderer):
+        message = []
+        if renderer.get("message"):
+            runs=renderer["message"].get("runs")
+            if runs:
+                for r in runs:
+                    if r:
+                        if r.get('emoji'):
+                            message.append(r['emoji']['image']['thumbnails'][1].get('url'))
+                        else:
+                            message.append(r.get('text',''))
+        return message
+
     def get_badges(self,renderer):
         isVerified = False
         isChatOwner = False

pytchat/processors/default/renderer/currency.py

@@ -23,7 +23,7 @@ symbols = {
     "PLN\xa0": {"fxtext": "PLN", "jptext": "ポーランド・ズロチ"},
     "R$": {"fxtext": "BRL", "jptext": "ブラジル・レアル"},
     "RUB\xa0": {"fxtext": "RUB", "jptext": "ロシア・ルーブル"},
-    "SEK\xa0": {"fxtext": "SEK", "jptext": "スウェーデン・クローネ"},
+    "SEK\xa0": {"fxtext": "SEK", "jptext": "スウェーデン・クローナ"},
     "£": {"fxtext": "GBP", "jptext": "英・ポンド"},
     "₩": {"fxtext": "KRW", "jptext": "韓国・ウォン"},
     "€": {"fxtext": "EUR", "jptext": "欧・ユーロ"},

pytchat/processors/default/renderer/paidsticker.py

@@ -1,12 +1,42 @@
 import re
 from . import currency
-from .paidmessage import LiveChatPaidMessageRenderer
+from .base import BaseRenderer
+superchat_regex = re.compile(r"^(\D*)(\d{1,3}(,\d{3})*(\.\d*)*\b)$")
 
-class LiveChatPaidStickerRenderer(LiveChatPaidMessageRenderer):
+class LiveChatPaidStickerRenderer(BaseRenderer):
     def __init__(self, item):
         super().__init__(item, "superSticker")
 
 
+    def get_snippet(self):
+        super().get_snippet()
+
+        self.author.name = self.renderer["authorName"]["simpleText"]
+
+        amountDisplayString, symbol, amount =(
+            self.get_amountdata(self.renderer)
+        )
+        self.message = ""
+        self.amountValue = amount
+        self.amountString = amountDisplayString
+        self.currency = currency.symbols[symbol]["fxtext"] if currency.symbols.get(symbol) else symbol
+        self.bgColor = self.renderer.get("moneyChipBackgroundColor", 0)
+        self.sticker = "https:"+self.renderer["sticker"]["thumbnails"][0]["url"]
+            
+
+
+    def get_amountdata(self,renderer):
+        amountDisplayString = renderer["purchaseAmountText"]["simpleText"]
+        m = superchat_regex.search(amountDisplayString)
+        if m:
+            symbol = m.group(1)
+            amount = float(m.group(2).replace(',',''))
+        else:
+            symbol = ""
+            amount = 0.0
+        return amountDisplayString, symbol, amount
+
+
  
 
  

pytchat/processors/simple_display_processor.py

@@ -31,7 +31,7 @@ class SimpleDisplayProcessor(ChatProcessor):
                     purchase_amount_text = ''
                 else:
                     root = ( action['addChatItemAction']['item'].get('liveChatPaidMessageRenderer') or 
-                            action['addChatItemAction']['item'].get('liveChatPaidMessageRenderer') )
+                            action['addChatItemAction']['item'].get('liveChatPaidStickerRenderer') )
                     if root:
                         author_name = root['authorName']['simpleText']
                         message = self._parse_message(root.get('message'))

pytchat/util/__init__.py

@@ -1,7 +1,7 @@
 import requests,json,datetime
 from .. import config
 
-def download(cls,url):
+def download(url):
     _session = requests.Session()
     html = _session.get(url, headers=config.headers)
     with open(str(datetime.datetime.now().strftime('%Y-%m-%d %H-%M-%S')
@@ -9,7 +9,7 @@ def download(cls,url):
         json.dump(html.json(),f,ensure_ascii=False)
 
 
-def save(cls,data,filename):
+def save(data,filename):
     with open(str(datetime.datetime.now().strftime('%Y-%m-%d %H-%M-%S')
                             )+filename,mode ='w',encoding='utf-8') as f:
         f.writelines(data)

requirements.txt

@@ -1,9 +1,4 @@
-aiohttp==3.6.0
-aioresponses==0.6.0
-mock==3.0.5
-mocker==1.1.1
-pytest==5.1.2
-pytest-mock==1.10.4
-pytz==2019.2
-requests==2.22.0
-urllib3==1.25.3
+aiohttp
+pytz
+requests
+urllib3

requirements_test.txt

@@ -0,0 +1,5 @@
+aioresponses
+mock
+mocker
+pytest
+pytest-mock

setup.py

@@ -1,14 +1,18 @@
-from setuptools import setup, find_packages
+from setuptools import setup, find_packages, Command
 from codecs import open
-from os import path
+from os import path, system
 import re
 
 package_name = "pytchat"
 
 root_dir = path.abspath(path.dirname(__file__))
 
-def _requires_from_file(filename):
-    return open(filename).read().splitlines()
+def _requirements():
+    return [name.rstrip() for name in open(path.join(root_dir, 'requirements.txt')).readlines()]
+
+def _test_requirements():
+    return [name.rstrip() for name in open(path.join(root_dir, 'requirements_test.txt')).readlines()]
+
 
 with open(path.join(root_dir, package_name, '__init__.py')) as f:
     init_text = f.read()
@@ -27,18 +31,20 @@ assert url
 with open('README.md', encoding='utf-8') as f:
     long_description = f.read()
 
+
+
 setup(
     name=package_name,
     packages=find_packages(),
-
     version=version,
-
     url=url,
     author=author,
     author_email=author_email,
     long_description=long_description,
     long_description_content_type='text/markdown',
     license=license,
+    install_requires=_requirements(),
+    tests_require=_test_requirements(),
     description="a python library for fetching youtube live chat.",
     classifiers=[
         'Natural Language :: Japanese',
@@ -52,5 +58,5 @@ setup(
         'License :: OSI Approved :: MIT License',
     ],
     keywords='youtube livechat asyncio',
-    install_requires=_requires_from_file('requirements.txt')
+
 )

tests/test_arcparam.py

@@ -0,0 +1,26 @@
+import pytest
+from pytchat.parser.replay import Parser
+import pytchat.config as config
+import requests, json
+from pytchat.paramgen import arcparam
+
+def test_arcparam_0(mocker):
+    param = arcparam.get("01234567890")
+    assert "op2w0wRyGjxDZzhhRFFvTE1ERXlNelExTmpjNE9UQWFFLXFvM2JrQkRRb0xNREV5TXpRMU5qYzRPVEFnQVElM0QlM0QoATAAOABAAEgEUhwIABAAGAAgACoOc3RhdGljY2hlY2tzdW1AAFgDYAFoAXIECAEQAXgA" == param
+
+
+def test_arcparam_1(mocker):
+    param = arcparam.get("01234567890", seektime = 100000)
+    assert "op2w0wR3GjxDZzhhRFFvTE1ERXlNelExTmpjNE9UQWFFLXFvM2JrQkRRb0xNREV5TXpRMU5qYzRPVEFnQVElM0QlM0QogNDbw_QCMAA4AEAASANSHAgAEAAYACAAKg5zdGF0aWNjaGVja3N1bUAAWANgAWgBcgQIARABeAA%3D" == param
+
+def test_arcparam_2(mocker):
+    param = arcparam.get("SsjCnHOk-Sk")
+    url=f"https://www.youtube.com/live_chat_replay/get_live_chat_replay?continuation={param}&pbj=1"
+    resp = requests.Session().get(url,headers = config.headers)
+    jsn = json.loads(resp.text)
+    parser = Parser()
+    _ , chatdata = parser.parse(jsn)
+    test_id = chatdata[0]["addChatItemAction"]["item"]["liveChatTextMessageRenderer"]["id"]
+    print(test_id)
+    assert "CjoKGkNMYXBzZTdudHVVQ0Zjc0IxZ0FkTnFnQjVREhxDSnlBNHV2bnR1VUNGV0dnd2dvZDd3NE5aZy0w" == test_id
+    

tests/test_compatible_processor.py

@@ -1,7 +1,7 @@
 import json
 import pytest
 import asyncio,aiohttp
-from pytchat.core_async.parser import Parser
+from pytchat.parser.live import Parser
 from pytchat.processors.compatible.processor import CompatibleProcessor
 from pytchat.exceptions import (
     NoLivechatRendererException,NoYtinitialdataException,
@@ -12,13 +12,15 @@ from pytchat.processors.compatible.renderer.paidmessage import LiveChatPaidMessa
 from pytchat.processors.compatible.renderer.paidsticker import LiveChatPaidStickerRenderer
 from pytchat.processors.compatible.renderer.legacypaid import LiveChatLegacyPaidMessageRenderer
 
+parser = Parser()
+
 def test_textmessage(mocker):
     '''api互換processorのテスト：通常テキストメッセージ'''
     processor = CompatibleProcessor()
 
     _json = _open_file("tests/testdata/compatible/textmessage.json")
 
-    _, chatdata = Parser.parse(json.loads(_json))
+    _, chatdata = parser.parse(json.loads(_json))
     data = {
         "video_id" : "",
         "timeout" : 7,
@@ -55,7 +57,7 @@ def test_newsponcer(mocker):
 
     _json = _open_file("tests/testdata/compatible/newSponsor.json")
 
-    _, chatdata = Parser.parse(json.loads(_json))
+    _, chatdata = parser.parse(json.loads(_json))
     data = {
         "video_id" : "",
         "timeout" : 7,
@@ -91,7 +93,7 @@ def test_superchat(mocker):
 
     _json = _open_file("tests/testdata/compatible/superchat.json")
 
-    _, chatdata = Parser.parse(json.loads(_json))
+    _, chatdata = parser.parse(json.loads(_json))
     data = {
         "video_id" : "",
         "timeout" : 7,

tests/test_livechat.py

@@ -1,6 +1,6 @@
 import pytest
-from pytchat.core_async.parser import Parser as AsyncParser
-from pytchat.core_multithread.parser import Parser as ThreadParser
+from pytchat.parser.live import Parser
+
 import json
 import asyncio,aiohttp
 

tests/test_liveparam.py

@@ -0,0 +1,9 @@
+import pytest
+from pytchat.paramgen import liveparam
+
+def test_liveparam_0(mocker):
+    _ts1= 1546268400
+    param = liveparam._build("01234567890",
+    *([_ts1*1000000 for i in range(5)]))
+    test_param="0ofMyAPiARp8Q2c4S0RRb0xNREV5TXpRMU5qYzRPVEFhUTZxNXdiMEJQUW83YUhSMGNITTZMeTkzZDNjdWVXOTFkSFZpWlM1amIyMHZiR2wyWlY5amFHRjBQM1k5TURFeU16UTFOamM0T1RBbWFYTmZjRzl3YjNWMFBURWdBZyUzRCUzRCiAuNbVqsrfAjAAOABAAkorCAEQABgAIAAqDnN0YXRpY2NoZWNrc3VtOgBAAEoCCAFQgLjW1arK3wJYA1CAuNbVqsrfAliAuNbVqsrfAmgBggEECAEQAIgBAKABgLjW1arK3wI%3D"
+    assert test_param == param

tests/test_parser.py

@@ -1,5 +1,5 @@
 import pytest
-from pytchat.core_async.parser import Parser
+from pytchat.parser.live import Parser
 import json
 import asyncio,aiohttp
 from aioresponses import aioresponses
@@ -11,7 +11,7 @@ from pytchat.exceptions import (
 def _open_file(path):
     with open(path,mode ='r',encoding = 'utf-8') as f:
         return f.read()
-
+parser = Parser()
 
 @aioresponses()
 def test_finishedlive(*mock):
@@ -21,7 +21,7 @@ def test_finishedlive(*mock):
     _text = json.loads(_text)
 
     try:    
-        Parser.parse(_text)
+        parser.parse(_text)
         assert False
     except NoContentsException:
         assert True
@@ -34,7 +34,7 @@ def test_parsejson(*mock):
     _text = json.loads(_text)
 
     try:    
-        Parser.parse(_text)
+        parser.parse(_text)
         jsn = _text
         timeout = jsn["response"]["continuationContents"]["liveChatContinuation"]["continuations"][0]["timedContinuationData"]["timeoutMs"]
         continuation = jsn["response"]["continuationContents"]["liveChatContinuation"]["continuations"][0]["timedContinuationData"]["continuation"]