From 350d3727d5cf7a9439cc7e150c9cc49f45f8f21b Mon Sep 17 00:00:00 2001
From: Yuanle Song <sylecn@gmail.com>
Date: Sun, 3 Mar 2019 17:14:22 +0800
Subject: [PATCH] v0.3.1 drop http link in local m3u8 file.

---
 m3u8downloader/__init__.py  |  2 +-
 m3u8downloader/main.py      | 32 ++++++++++++++++++++++++++++++++
 m3u8downloader/test_main.py |  6 ++++++
 operational                 | 25 +++++++++++++++++++++++++
 requirements.txt            |  2 +-
 5 files changed, 65 insertions(+), 2 deletions(-)

diff --git a/m3u8downloader/__init__.py b/m3u8downloader/__init__.py
index 493f741..260c070 100644
--- a/m3u8downloader/__init__.py
+++ b/m3u8downloader/__init__.py
@@ -1 +1 @@
-__version__ = "0.3.0"
+__version__ = "0.3.1"
diff --git a/m3u8downloader/main.py b/m3u8downloader/main.py
index 0f012c4..1ea4004 100644
--- a/m3u8downloader/main.py
+++ b/m3u8downloader/main.py
@@ -96,6 +96,37 @@ def get_fullpath(filename):
     return os.path.abspath(os.path.expandvars(os.path.expanduser(filename)))
 
 
+def http_line_to_relpath_line(url_line):
+    """convert a url line in m3u8 file to use relative path in local fs.
+
+    """
+    r = urlparse(url_line)
+    return r.path[1:]
+
+
+def drop_http_link_in_m3u8_file(local_m3u8_filename):
+    """drop http protocol, host part for all resource lines.
+
+    """
+    with open(local_m3u8_filename, 'r') as f:
+        content = f.read()
+    if 'http' not in content:
+        logger.info("media playlist m3u8 file doesn't contain http link")
+        return
+    with open(local_m3u8_filename, 'w') as f:
+        for line in content.split('\n'):
+            if line.startswith('#'):
+                f.write(line)
+                f.write('\n')
+            if line.strip() == '':
+                f.write(line)
+                f.write('\n')
+            if line.startswith('http'):
+                f.write(http_line_to_relpath_line(line))
+                f.write('\n')
+    logger.info("updated media playlist m3u8 file: %s", local_m3u8_filename)
+
+
 class M3u8Downloader:
     def __init__(self, url, output_filename, tempdir="."):
         self.start_url = url
@@ -223,6 +254,7 @@ class M3u8Downloader:
 
         """
         self.media_playlist_localfile = self.mirror_url_resource(url)
+        drop_http_link_in_m3u8_file(self.media_playlist_localfile)
         if content is None:
             content = get_url_content(url)
 
diff --git a/m3u8downloader/test_main.py b/m3u8downloader/test_main.py
index 74feebd..7029b42 100644
--- a/m3u8downloader/test_main.py
+++ b/m3u8downloader/test_main.py
@@ -12,8 +12,14 @@ from m3u8downloader.main import is_higher_resolution
 from m3u8downloader.main import get_url_path
 from m3u8downloader.main import get_basename
 from m3u8downloader.main import get_fullpath
+from m3u8downloader.main import http_line_to_relpath_line
 
 
+# test for http_line_to_relpath_line
+def test_http_line_to_relpath_line():
+    assert http_line_to_relpath_line("https://m3u.x8sinablog.net/20180627/1530068509/vts/audio000.png") == "20180627/1530068509/vts/audio000.png"
+    assert http_line_to_relpath_line("https://m3u.x8sinablog.net/20180627/1530068509/vts/audio000.png?abc=def") == "20180627/1530068509/vts/audio000.png"
+
 # test for join
 def test_join():
     assert os.path.normpath(os.path.join("/foo/bar/baz", "./abc.txt")) == "/foo/bar/baz/abc.txt"
diff --git a/operational b/operational
index 8bbb880..53df504 100644
--- a/operational
+++ b/operational
@@ -39,6 +39,31 @@ to handle it on ryzen5.
 
 * current                                                             :entry:
 ** 
+** 2019-03-03 generate mp4 for audio sequence fail.
+probably because of png ext.
+https://m3u.x8sinablog.net/20180627/1530068509/vts/audio000.png
+
+ffprobe on the png file show correct result.
+it's aac audio.
+
+- try again.
+
+  ffmpeg -allowed_extensions ALL -i "/home/sylecn/.cache/m3u8downloader/28】/apiv286.m3u8" -acodec copy -vcodec copy -bsf:a aac_adtstoasc ~/d/t2/foo.m4a
+
+  it's not the file extension's problem.
+  it's the URL link.
+
+- audio .m3u8 uses absolute link that is not hosted on the same domain.
+
+  media playlist:
+  ~/.cache/m3u8downloader/28】/apiv286.m3u8
+  https://m3u.x8sinablog.net/20180627/1530068509/vts/audio000.png
+
+  I need to truncate the domain part for the local m3u8 files.
+  plain mirror is not enough.
+
+  it works.
+
 ** 2019-03-03 add progress tracking log.
 do a commit before I add this.
 
diff --git a/requirements.txt b/requirements.txt
index 4824c83..5d398c4 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,4 @@
 wells==1.4.1
-requests
+requests==2.21.0
 # six==1.12.0
 # raven==6.0.0
-- 
GitLab