From 3936310e7e9bbd5194bc07403f775c4a50dfb16b Mon Sep 17 00:00:00 2001
From: James Taylor <user234683@users.noreply.github.com>
Date: Sat, 21 Dec 2019 11:58:02 -0800
Subject: Fix extract_approx_int. Fixes incorrect subscriber count on channels.
It wasn't working because decimals such as 15.1M weren't considered, so it
was extracting "1M"
---
youtube/yt_data_extract/common.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
(limited to 'youtube')
diff --git a/youtube/yt_data_extract/common.py b/youtube/yt_data_extract/common.py
index dfeae64..221b523 100644
--- a/youtube/yt_data_extract/common.py
+++ b/youtube/yt_data_extract/common.py
@@ -142,12 +142,12 @@ def extract_int(string, default=None):
return default
def extract_approx_int(string):
- '''e.g. "15M" from "15M subscribers"'''
+ '''e.g. "15.1M" from "15.1M subscribers"'''
if not isinstance(string, str):
string = extract_str(string)
if not string:
return None
- match = re.search(r'(\d+[KMBTkmbt])', string.replace(',', ''))
+ match = re.search(r'(\d+(?:\.\d+)?[KMBTkmbt])', string.replace(',', ''))
if match is None:
return None
return match.group(1)
--
cgit v1.2.3