From e2bae2efd5836f33594f800953552dea3509864d Mon Sep 17 00:00:00 2001 From: blank X Date: Thu, 24 Dec 2020 22:09:02 +0700 Subject: [PATCH] More fixes --- sukuinote/__init__.py | 11 ++++++ sukuinote/plugins/saucenao.py | 68 +++++++++++++++++++---------------- 2 files changed, 48 insertions(+), 31 deletions(-) diff --git a/sukuinote/__init__.py b/sukuinote/__init__.py index 7f5eab2..f4dbf9c 100644 --- a/sukuinote/__init__.py +++ b/sukuinote/__init__.py @@ -190,3 +190,14 @@ async def get_file_mimetype(filename): stdout, _ = await proc.communicate() mimetype = stdout.decode().strip() return mimetype or '' + +async def get_file_ext(filename): + proc = await asyncio.create_subprocess_exec('file', '--brief', '--extension', filename, stdout=asyncio.subprocess.PIPE) + stdout, _ = await proc.communicate() + ext = stdout.decode().strip().split('/', maxsplit=1)[0] + if not ext or ext == '???': + mimetype = await get_file_mimetype(filename) + ext = mimetypes.guess_extension(mimetype) or '.bin' + if not ext.startswith('.'): + ext = '.' + ext + return ext diff --git a/sukuinote/plugins/saucenao.py b/sukuinote/plugins/saucenao.py index 25a696f..0b65d32 100644 --- a/sukuinote/plugins/saucenao.py +++ b/sukuinote/plugins/saucenao.py @@ -6,7 +6,7 @@ from decimal import Decimal from urllib.parse import urlparse, urlunparse, quote as urlencode from bs4 import BeautifulSoup from pyrogram import Client, filters -from .. import config, help_dict, log_errors, public_log_errors, session, get_file_mimetype, progress_callback +from .. import config, help_dict, log_errors, public_log_errors, session, get_file_mimetype, progress_callback, get_file_ext async def download_file(url, filename): async with session.get(url) as resp: @@ -44,7 +44,7 @@ async def saucenao(client, message): await proc.communicate() filename = new_path with open(filename, 'rb') as file: - async with session.post(f'https://saucenao.com/search.php?db=999&output_type=2&numres=5&api_key={urlencode(config["config"]["saucenao_api"])}', data={'file': file}) as resp: + async with session.post(f'https://saucenao.com/search.php?db=999&output_type=2&api_key={urlencode(config["config"]["saucenao_api"])}', data={'file': file}) as resp: json = await resp.json() if json['header']['status']: await reply.edit_text(f'{json["header"]["status"]}: {html.escape(json["header"].get("message", "No message"))}') @@ -54,44 +54,50 @@ async def saucenao(client, message): to_image = False filename = os.path.join(tempdir, '0') for result in json['results']: + if not result['data'].get('ext_urls'): + continue atext = f'{html.escape(result["header"]["index_name"])}' if Decimal(result['header']['similarity']) < minimum_similarity: atext += ' (low similarity result)' atext += '' - if result['data'].get('ext_urls'): - atext += '\nURL' - if len(result['data']['ext_urls']) > 1: - atext += 's:\n' - atext += '\n'.join(map(html.escape, result['data']['ext_urls'])) - else: - atext += f': {html.escape(result["data"]["ext_urls"][0])}' - if not to_image: - for url in result['data']['ext_urls']: - if await download_file(url, filename): - with open(filename) as file: - soup = BeautifulSoup(file.read()) - pimg = soup.find(lambda tag: tag.name == 'meta' and tag.attrs.get('property') == 'og:image' and tag.attrs.get('content')) + atext += '\nURL' + if len(result['data']['ext_urls']) > 1: + atext += 's:\n' + atext += '\n'.join(map(html.escape, result['data']['ext_urls'])) + else: + atext += f': {html.escape(result["data"]["ext_urls"][0])}' + if not to_image: + for url in result['data']['ext_urls']: + if await download_file(url, filename): + with open(filename) as file: + soup = BeautifulSoup(file.read()) + pimg = soup.find(lambda tag: tag.name == 'meta' and tag.attrs.get('property') == 'og:image' and tag.attrs.get('content')) + if pimg: + pimg = pimg.attrs.get('content', '').strip() if pimg: - pimg = pimg.attrs.get('content', '').strip() - if pimg: - parsed = list(urlparse(pimg)) - if not parsed[0]: - parsed[0] = 'https' - pimg = urlunparse(parsed) - if parsed[0] not in ('http', 'https'): - continue - if await download_file(pimg, filename): - to_image = True - break - else: - await download_file(result['header']['thumbnail'], filename) - to_image = True + parsed = list(urlparse(pimg)) + if not parsed[0]: + parsed[0] = 'https' + pimg = urlunparse(parsed) + if parsed[0] not in ('http', 'https'): + continue + if await download_file(pimg, filename): + to_image = True + break + else: + await download_file(result['header']['thumbnail'], filename) + to_image = True atext += '\n\n' - if len((await client.parser.parse(caption + atext, 'html'))['message']) <= 1024: + length = len((await client.parser.parse(caption + atext, 'html'))['message']) + if length <= 1024: caption += atext + if length > 4096: + break text += atext try: - await message.reply_photo(filename, caption=caption) + ext = await get_file_ext(filename) + os.rename(filename, filename + ext) + await message.reply_photo(filename + ext, caption=caption) except Exception: await reply.edit_text(text) else: