Add test

pytchat/tool/asyncdl.py

@@ -13,12 +13,23 @@ headers = config.headers
 REPLAY_URL = "https://www.youtube.com/live_chat_replay/" \
              "get_live_chat_replay?continuation="
 
-def _divide(start, end, count):
-    min_interval = 120
-    if (not isinstance(start,int) or 
-        not isinstance(end,int) or 
-        not isinstance(count,int)):
-        raise ValueError("start/end/count must be int")
+def _split(start, end, count, min_interval = 120):
+    """
+    Split section from `start` to `end` into `count` pieces,
+    and returns the beginning of each piece. 
+    The `count` is adjusted so that the length of each piece
+    is no smaller than `min_interval`.
+
+    Returns:
+    --------
+    List of the beginning position of each piece.
+    """
+    
+    if not (isinstance(start,int) or isinstance(start,float)) or \
+       not (isinstance(end,int) or isinstance(end,float)):
+        raise ValueError("start/end must be int or float")
+    if not isinstance(count,int):
+        raise ValueError("count must be int")
     if start>end:
         raise ValueError("end must be equal to or greater than start.")
     if count<1:
@@ -39,7 +50,7 @@ def ready_blocks(video_id, duration, div, callback):
     async def _get_blocks( video_id, duration, div, callback):
         async with aiohttp.ClientSession() as session:
             futures = [_create_block(session, video_id, pos, seektime, callback)
-                for pos, seektime in enumerate(_divide(-1, duration, div))]
+                for pos, seektime in enumerate(_split(-1, duration, div))]
             return await asyncio.gather(*futures,return_exceptions=True)
 
     async def _create_block(session, video_id, pos, seektime, callback):

pytchat/tool/block.py

@@ -3,27 +3,27 @@ class Block:
 
     Parameter:
     ---------
-    pos : int
+    pos : int :
         index of this block on block list.
 
-    first : int
+    first : int :
         videoOffsetTimeMs of chat_data[0]
         
-    last : int
-        videoOffsetTimeMs of the last chat_data current read.
+    last : int :
+        videoOffsetTimeMs of the last chat_data.
         (chat_data[-1])
 
         this value increases as fetching chatdata progresses.
 
-    temp_last : int
-        temporary videoOffsetTimeMs of last chat data,
+    temp_last : int :
+        target videoOffsetTimeMs of last chat data for download,
         equals to first videoOffsetTimeMs of next block.
-        when download worker reaches this offset, the download will stop.
+        when download worker reaches this offset, stop downloading.
 
-    continuation : str
+    continuation : str :
         continuation param of last chat data.
 
-    chat_data : List
+    chat_data : List 
     """
     def __init__(self, pos=0, first=0, last=0,
                 continuation='', chat_data=[]):

pytchat/tool/dlworker.py

@@ -2,15 +2,15 @@ from . import parser
 
 class DownloadWorker:
     """
-    DownloadWorker : associates a download session with a block.
+    DownloadWorker associates a download session with a block.
 
     Parameter
     ----------
-    fetch : func
+    fetch :
         download function of asyncdl
 
-    block : Block
-        chunk of chat_data
+    block :
+        Block object that includes chat_data
     """
     def __init__(self, fetch, block):
         self.block = block

pytchat/tool/downloader.py

@@ -55,8 +55,9 @@ class Downloader:
         return (
             self.ready_blocks()
                 .remove_duplicate_head()
-                .set_temporary_last()
+#                .set_temporary_last()
                 .remove_overwrap()
+                .set_temporary_last()
                 .download_blocks()
                 .remove_duplicate_tail()
                 .combine()

pytchat/tool/duplcheck.py

@@ -52,7 +52,6 @@ def duplicate_head(blocks):
                 and
             type_0 == type_1
         )
-
     ret = [blocks[i] for i in range(len(blocks)-1)
         if (len(blocks[i].chat_data)>0 and 
         not is_duplicate_head(i) )]
@@ -75,9 +74,9 @@ def duplicate_tail(blocks):
             type_0 == type_1
         )
 
-    ret = [blocks[i] for i in range(0,len(blocks)-1)
+    ret = [blocks[i] for i in range(0,len(blocks))
         if i == 0 or not  is_duplicate_tail(i) ]
-    ret.append(blocks[-1])
+    #ret.append(blocks[-1])
     return ret
 
 def overwrap(blocks):
@@ -103,3 +102,9 @@ def overwrap(blocks):
         b = a+1
     ret.append(blocks[-1])
     return ret
+
+def _dump(blocks):
+    print(__name__)
+    print(f"----------        first         last   temp_last {'':>3}---")
+    for i,block in enumerate(blocks):
+        print(f"block[{i:3}]   {block.first:>10}   {block.last:>10}  {block.temp_last:>10}")