diff --git a/CHANGELOG.rst b/CHANGELOG.rst index c38c312..ec9a51f 100644 --- a/CHANGELOG.rst +++ b/CHANGELOG.rst @@ -16,6 +16,7 @@ v2.1.0 (IN PROGRESS) * Added pagination iteraton via `pagination_iterator` (Thanks @FredericoCeratto for the suggestion) * Added a way to get pagination info out of lists that is slightly less digging-around-in-internals via `get_pagination_info` (Thanks @s427 for the inciting report) * Added missing `replies_policy` and `exclusive` parameters to list creation and update methods. +* Add status length counter `get_status_length` (Thanks @yuletide for the suggestion) v2.0.1 ------ diff --git a/docs/12_utilities.rst b/docs/12_utilities.rst index 5a35bfe..435c937 100644 --- a/docs/12_utilities.rst +++ b/docs/12_utilities.rst @@ -33,3 +33,4 @@ Cache control Other utilities --------------- .. automethod:: Mastodon.get_approx_server_time +.. automethod:: Mastodon.get_status_length diff --git a/mastodon/_url_regex.py b/mastodon/_url_regex.py new file mode 100644 index 0000000..7275fc9 --- /dev/null +++ b/mastodon/_url_regex.py @@ -0,0 +1,6 @@ +import json +import re + +URL_REGEX_JSON = r"""{"source":"(((?:[^A-Za-z0-9@＠$#＃\\uFFFE\\uFEFF\\uFFFF]|[\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069]|^))((https?:\\/\\/)((?:(?:(?:[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/](?:[_-]|[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/])*)?[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/]\\.)*(?:(?:[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/](?:-|[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/])*)?[^/\\!'#%&'\\(\\)*\\+,\\\\\\-\\.\\/:;<=>\\?@\\[\\]\\^_{|}~\\$//\\x09-\\x0D\\x20\\x85\\xA0\\u1680\\u180E\\u2000-\\u200A\\u2028\\u2029\\u202F\\u205F\\u3000//\\uFFFE\\uFEFF\\uFFFF//\\u202A-\\u202E\\u061C\\u200E\\u200F\\u2066\\u2067\\u2068\\u2069/]\\.)(?:(?:(?:삼성|닷컴|닷넷|香格里拉|餐厅|食品|飞利浦|電訊盈科|集团|通販|购物|谷歌|诺基亚|联通|网络|网站|网店|网址|组织机构|移动|珠宝|点看|游戏|淡马锡|机构|書籍|时尚|新闻|政府|政务|招聘|手表|手机|我爱你|慈善|微博|广东|工行|家電|娱乐|天主教|大拿|大众汽车|在线|嘉里大酒店|嘉里|商标|商店|商城|公益|公司|八卦|健康|信息|佛山|企业|中文网|中信|世界|ポイント|ファッション|セール|ストア|コム|グーグル|クラウド|みんな|คอม|संगठन|नेट|कॉम|همراه|موقع|موبايلي|كوم|كاثوليك|عرب|شبكة|بيتك|بازار|العليان|ارامكو|اتصالات|ابوظبي|קום|сайт|рус|орг|онлайн|москва|ком|католик|дети|zuerich|zone|zippo|zip|zero|zara|zappos|yun|youtube|you|yokohama|yoga|yodobashi|yandex|yamaxun|yahoo|yachts|xyz|xxx|xperia|xin|xihuan|xfinity|xerox|xbox|wtf|wtc|wow|world|works|work|woodside|wolterskluwer|wme|winners|wine|windows|win|williamhill|wiki|wien|whoswho|weir|weibo|wedding|wed|website|weber|webcam|weatherchannel|weather|watches|watch|warman|wanggou|wang|walter|walmart|wales|vuelos|voyage|voto|voting|vote|volvo|volkswagen|vodka|vlaanderen|vivo|viva|vistaprint|vista|vision|visa|virgin|vip|vin|villas|viking|vig|video|viajes|vet|versicherung|vermögensberatung|vermögensberater|verisign|ventures|vegas|vanguard|vana|vacations|ups|uol|uno|university|unicom|uconnect|ubs|ubank|tvs|tushu|tunes|tui|tube|trv|trust|travelersinsurance|travelers|travelchannel|travel|training|trading|trade|toys|toyota|town|tours|total|toshiba|toray|top|tools|tokyo|today|tmall|tkmaxx|tjx|tjmaxx|tirol|tires|tips|tiffany|tienda|tickets|tiaa|theatre|theater|thd|teva|tennis|temasek|telefonica|telecity|tel|technology|tech|team|tdk|tci|taxi|tax|tattoo|tatar|tatamotors|target|taobao|talk|taipei|tab|systems|symantec|sydney|swiss|swiftcover|swatch|suzuki|surgery|surf|support|supply|supplies|sucks|style|study|studio|stream|store|storage|stockholm|stcgroup|stc|statoil|statefarm|statebank|starhub|star|staples|stada|srt|srl|spreadbetting|spot|sport|spiegel|space|soy|sony|song|solutions|solar|sohu|software|softbank|social|soccer|sncf|smile|smart|sling|skype|sky|skin|ski|site|singles|sina|silk|shriram|showtime|show|shouji|shopping|shop|shoes|shiksha|shia|shell|shaw|sharp|shangrila|sfr|sexy|sex|sew|seven|ses|services|sener|select|seek|security|secure|seat|search|scot|scor|scjohnson|science|schwarz|schule|school|scholarships|schmidt|schaeffler|scb|sca|sbs|sbi|saxo|save|sas|sarl|sapo|sap|sanofi|sandvikcoromant|sandvik|samsung|samsclub|salon|sale|sakura|safety|safe|saarland|ryukyu|rwe|run|ruhr|rugby|rsvp|room|rogers|rodeo|rocks|rocher|rmit|rip|rio|ril|rightathome|ricoh|richardli|rich|rexroth|reviews|review|restaurant|rest|republican|report|repair|rentals|rent|ren|reliance|reit|reisen|reise|rehab|redumbrella|redstone|red|recipes|realty|realtor|realestate|read|raid|radio|racing|qvc|quest|quebec|qpon|pwc|pub|prudential|pru|protection|property|properties|promo|progressive|prof|productions|prod|pro|prime|press|praxi|pramerica|post|porn|politie|poker|pohl|pnc|plus|plumbing|playstation|play|place|pizza|pioneer|pink|ping|pin|pid|pictures|pictet|pics|piaget|physio|photos|photography|photo|phone|philips|phd|pharmacy|pfizer|pet|pccw|pay|passagens|party|parts|partners|pars|paris|panerai|panasonic|pamperedchef|page|ovh|ott|otsuka|osaka|origins|orientexpress|organic|org|orange|oracle|open|ooo|onyourside|online|onl|ong|one|omega|ollo|oldnavy|olayangroup|olayan|okinawa|office|off|observer|obi|nyc|ntt|nrw|nra|nowtv|nowruz|now|norton|northwesternmutual|nokia|nissay|nissan|ninja|nikon|nike|nico|nhk|ngo|nfl|nexus|nextdirect|next|news|newholland|new|neustar|network|netflix|netbank|net|nec|nba|navy|natura|nationwide|name|nagoya|nadex|nab|mutuelle|mutual|museum|mtr|mtpc|mtn|msd|movistar|movie|mov|motorcycles|moto|moscow|mortgage|mormon|mopar|montblanc|monster|money|monash|mom|moi|moe|moda|mobily|mobile|mobi|mma|mls|mlb|mitsubishi|mit|mint|mini|mil|microsoft|miami|metlife|merckmsd|meo|menu|men|memorial|meme|melbourne|meet|media|med|mckinsey|mcdonalds|mcd|mba|mattel|maserati|marshalls|marriott|markets|marketing|market|map|mango|management|man|makeup|maison|maif|madrid|macys|luxury|luxe|lupin|lundbeck|ltda|ltd|lplfinancial|lpl|love|lotto|lotte|london|lol|loft|locus|locker|loans|loan|llp|llc|lixil|living|live|lipsy|link|linde|lincoln|limo|limited|lilly|like|lighting|lifestyle|lifeinsurance|life|lidl|liaison|lgbt|lexus|lego|legal|lefrak|leclerc|lease|lds|lawyer|law|latrobe|latino|lat|lasalle|lanxess|landrover|land|lancome|lancia|lancaster|lamer|lamborghini|ladbrokes|lacaixa|kyoto|kuokgroup|kred|krd|kpn|kpmg|kosher|komatsu|koeln|kiwi|kitchen|kindle|kinder|kim|kia|kfh|kerryproperties|kerrylogistics|kerryhotels|kddi|kaufen|juniper|juegos|jprs|jpmorgan|joy|jot|joburg|jobs|jnj|jmp|jll|jlc|jio|jewelry|jetzt|jeep|jcp|jcb|java|jaguar|iwc|iveco|itv|itau|istanbul|ist|ismaili|iselect|irish|ipiranga|investments|intuit|international|intel|int|insure|insurance|institute|ink|ing|info|infiniti|industries|inc|immobilien|immo|imdb|imamat|ikano|iinet|ifm|ieee|icu|ice|icbc|ibm|hyundai|hyatt|hughes|htc|hsbc|how|house|hotmail|hotels|hoteles|hot|hosting|host|hospital|horse|honeywell|honda|homesense|homes|homegoods|homedepot|holiday|holdings|hockey|hkt|hiv|hitachi|hisamitsu|hiphop|hgtv|hermes|here|helsinki|help|healthcare|health|hdfcbank|hdfc|hbo|haus|hangout|hamburg|hair|guru|guitars|guide|guge|gucci|guardian|group|grocery|gripe|green|gratis|graphics|grainger|gov|got|gop|google|goog|goodyear|goodhands|goo|golf|goldpoint|gold|godaddy|gmx|gmo|gmbh|gmail|globo|global|gle|glass|glade|giving|gives|gifts|gift|ggee|george|genting|gent|gea|gdn|gbiz|gay|garden|gap|games|game|gallup|gallo|gallery|gal|fyi|futbol|furniture|fund|fun|fujixerox|fujitsu|ftr|frontier|frontdoor|frogans|frl|fresenius|free|fox|foundation|forum|forsale|forex|ford|football|foodnetwork|food|foo|fly|flsmidth|flowers|florist|flir|flights|flickr|fitness|fit|fishing|fish|firmdale|firestone|fire|financial|finance|final|film|fido|fidelity|fiat|ferrero|ferrari|feedback|fedex|fast|fashion|farmers|farm|fans|fan|family|faith|fairwinds|fail|fage|extraspace|express|exposed|expert|exchange|everbank|events|eus|eurovision|etisalat|esurance|estate|esq|erni|ericsson|equipment|epson|epost|enterprises|engineering|engineer|energy|emerck|email|education|edu|edeka|eco|eat|earth|dvr|dvag|durban|dupont|duns|dunlop|duck|dubai|dtv|drive|download|dot|doosan|domains|doha|dog|dodge|doctor|docs|dnp|diy|dish|discover|discount|directory|direct|digital|diet|diamonds|dhl|dev|design|desi|dentist|dental|democrat|delta|deloitte|dell|delivery|degree|deals|dealer|deal|dds|dclk|day|datsun|dating|date|data|dance|dad|dabur|cyou|cymru|cuisinella|csc|cruises|cruise|crs|crown|cricket|creditunion|creditcard|credit|cpa|courses|coupons|coupon|country|corsica|coop|cool|cookingchannel|cooking|contractors|contact|consulting|construction|condos|comsec|computer|compare|company|community|commbank|comcast|com|cologne|college|coffee|codes|coach|clubmed|club|cloud|clothing|clinique|clinic|click|cleaning|claims|cityeats|city|citic|citi|citadel|cisco|circle|cipriani|church|chrysler|chrome|christmas|chloe|chintai|cheap|chat|chase|charity|channel|chanel|cfd|cfa|cern|ceo|center|ceb|cbs|cbre|cbn|cba|catholic|catering|cat|casino|cash|caseih|case|casa|cartier|cars|careers|career|care|cards|caravan|car|capitalone|capital|capetown|canon|cancerresearch|camp|camera|cam|calvinklein|call|cal|cafe|cab|bzh|buzz|buy|business|builders|build|bugatti|budapest|brussels|brother|broker|broadway|bridgestone|bradesco|box|boutique|bot|boston|bostik|bosch|boots|booking|book|boo|bond|bom|bofa|boehringer|boats|bnpparibas|bnl|bmw|bms|blue|bloomberg|blog|blockbuster|blanco|blackfriday|black|biz|bio|bingo|bing|bike|bid|bible|bharti|bet|bestbuy|best|berlin|bentley|beer|beauty|beats|bcn|bcg|bbva|bbt|bbc|bayern|bauhaus|basketball|baseball|bargains|barefoot|barclays|barclaycard|barcelona|bar|bank|band|bananarepublic|banamex|baidu|baby|azure|axa|aws|avianca|autos|auto|author|auspost|audio|audible|audi|auction|attorney|athleta|associates|asia|asda|arte|art|arpa|army|archi|aramco|arab|aquarelle|apple|app|apartments|aol|anz|anquan|android|analytics|amsterdam|amica|amfam|amex|americanfamily|americanexpress|alstom|alsace|ally|allstate|allfinanz|alipay|alibaba|alfaromeo|akdn|airtel|airforce|airbus|aigo|aig|agency|agakhan|africa|afl|afamilycompany|aetna|aero|aeg|adult|ads|adac|actor|active|aco|accountants|accountant|accenture|academy|abudhabi|abogado|able|abc|abbvie|abbott|abb|abarth|aarp|aaa|onion)(?=[^0-9a-zA-Z@+-]|$))|(?:(?:한국|香港|澳門|新加坡|台灣|台湾|中國|中国|გე|ລາວ|ไทย|ලංකා|ഭാരതം|ಭಾರತ|భారత్|சிங்கப்பூர்|இலங்கை|இந்தியா|ଭାରତ|ભારત|ਭਾਰਤ|ভাৰত|ভারত|বাংলা|भारोत|भारतम्|भारत|ڀارت|پاکستان|موريتانيا|مليسيا|مصر|قطر|فلسطين|عمان|عراق|سورية|سودان|تونس|بھارت|بارت|ایران|امارات|المغرب|السعودية|الجزائر|البحرين|الاردن|հայ|қаз|укр|срб|рф|мон|мкд|ею|бел|бг|ευ|ελ|zw|zm|za|yt|ye|ws|wf|vu|vn|vi|vg|ve|vc|va|uz|uy|us|um|uk|ug|ua|tz|tw|tv|tt|tr|tp|to|tn|tm|tl|tk|tj|th|tg|tf|td|tc|sz|sy|sx|sv|su|st|ss|sr|so|sn|sm|sl|sk|sj|si|sh|sg|se|sd|sc|sb|sa|rw|ru|rs|ro|re|qa|py|pw|pt|ps|pr|pn|pm|pl|pk|ph|pg|pf|pe|pa|om|nz|nu|nr|np|no|nl|ni|ng|nf|ne|nc|na|mz|my|mx|mw|mv|mu|mt|ms|mr|mq|mp|mo|mn|mm|ml|mk|mh|mg|mf|me|md|mc|ma|ly|lv|lu|lt|ls|lr|lk|li|lc|lb|la|kz|ky|kw|kr|kp|kn|km|ki|kh|kg|ke|jp|jo|jm|je|it|is|ir|iq|io|in|im|il|ie|id|hu|ht|hr|hn|hm|hk|gy|gw|gu|gt|gs|gr|gq|gp|gn|gm|gl|gi|gh|gg|gf|ge|gd|gb|ga|fr|fo|fm|fk|fj|fi|eu|et|es|er|eh|eg|ee|ec|dz|do|dm|dk|dj|de|cz|cy|cx|cw|cv|cu|cr|co|cn|cm|cl|ck|ci|ch|cg|cf|cd|cc|ca|bz|by|bw|bv|bt|bs|br|bq|bo|bn|bm|bl|bj|bi|bh|bg|bf|be|bd|bb|ba|az|ax|aw|au|at|as|ar|aq|ao|an|am|al|ai|ag|af|ae|ad|ac)(?=[^0-9a-zA-Z@+-]|$))|(?:xn--[\\-0-9a-z]+))))(?::([0-9]+))?(\\/(?:(?:[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*(?:\\((?:[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]+|(?:[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*\\([a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]+\\)[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*))\\)[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*)*[\\+\\-a-z\\u0400-\\u04FF0-9=_#\\/\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]|(?:\\((?:[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]+|(?:[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*\\([a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]+\\)[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]*))\\)))|(?:@[a-z\\u0400-\\u04FF0-9!\\*';:=\\+,\\.\\$\\/%#\\[\\]\\-\\u2013_~@\\|&\\xC0-\\xD6\\xD8-\\xF6\\xF8-\\xFF\\u0100-\\u024F\\u0253\\u0254\\u0256\\u0257\\u0259\\u025B\\u0263\\u0268\\u026F\\u0272\\u0289\\u028B\\u02BB\\u0300-\\u036F\\u1E00-\\u1EFF]+\\/))*)?(\\?[a-z0-9!?\\*'@\\(\\);:&=\\+\\$\\/%#\\[\\]\\-_\\.,~|]*[a-z0-9\\-_&=#\\/])?))","flags":"gi"}""" +URL_REGEX_SOURCE = json.loads(URL_REGEX_JSON)["source"] +url_regex = re.compile(URL_REGEX_SOURCE, re.IGNORECASE) \ No newline at end of file diff --git a/mastodon/compat.py b/mastodon/compat.py index acbb3cf..1112874 100644 --- a/mastodon/compat.py +++ b/mastodon/compat.py @@ -45,3 +45,9 @@ except: class Path: pass +IMPL_HAS_GRAPHEME = True +try: + import grapheme +except: + IMPL_HAS_GRAPHEME = False + grapheme = None diff --git a/mastodon/utility.py b/mastodon/utility.py index e218db9..ca4ae09 100644 --- a/mastodon/utility.py +++ b/mastodon/utility.py @@ -7,7 +7,7 @@ import copy import warnings from mastodon.errors import MastodonAPIError, MastodonIllegalArgumentError, MastodonNotFoundError, MastodonVersionError -from mastodon.compat import IMPL_HAS_BLURHASH, blurhash +from mastodon.compat import IMPL_HAS_BLURHASH, blurhash, IMPL_HAS_GRAPHEME, grapheme from mastodon.internals import Mastodon as Internals from mastodon.versions import parse_version_string, max_version, api_version @@ -16,8 +16,8 @@ from typing import Optional, Union, Dict, Iterator from mastodon.return_types import PaginatableList, PaginationInfo, PaginatableList from mastodon.types_base import Entity, try_cast -# Class level: - +from ._url_regex import url_regex +import unicodedata class Mastodon(Internals): def set_language(self, lang): @@ -320,3 +320,30 @@ class Mastodon(Internals): current_page = self.fetch_next(current_page) else: current_page = self.fetch_previous(current_page) + + @staticmethod + def get_status_length(text: str, spoiler_text: str = "") -> int: + """ + For a given status `text` and `spoiler_text`, return how many characters this status counts as + when computing the status length and comparing it against the limit. + + Note that there are other limits you may run into, such as the maximum length of a URL, or the + maximum length of a usernames domain part. But as long as you do *normal* things, this function + will return the correct length for the status text. + """ + if not IMPL_HAS_GRAPHEME: + raise NotImplementedError( + 'To use the get_status_length function, please install the grapheme Python module.') + + username_regex = re.compile(r'(^|[^/\w])@(([a-z0-9_]+)@[a-z0-9\.\-]+[a-z0-9]+)', re.IGNORECASE) + + def countable_text(input_text: str) -> str: + # Transform text such that it has the correct length for counting + # post text lengths against the limit + def _url_repl(m: re.Match) -> str: + return m.group(2) + ("x" * 23) + text = url_regex.sub(_url_repl, input_text) + text = username_regex.sub(r'\1@\3', text) + return text + + return grapheme.length(countable_text(text)) + grapheme.length(spoiler_text) diff --git a/tests/test_status_length.py b/tests/test_status_length.py new file mode 100644 index 0000000..a7059e2 --- /dev/null +++ b/tests/test_status_length.py @@ -0,0 +1,58 @@ + +import pytest + +from mastodon import Mastodon + +TEST_CASES = [ + # Simple + ("", 0), + ("hello", 5), + (" leading and trailing spaces ", 31), + (" tabs\tand\nnewlines\r\n", 19), + + # URLs - schemes, TLDs, IPv4/IPv6, ports, creds + ("check http://example.com and https://example.org/page?x=1#frag", 1000 - 943), + ("ftp://files.example.net/resource", 1000 - 968), + ("http://user:pass@example.com:8080/path", 1000 - 962), + ("http://127.0.0.1:3000/health", 1000 - 972), + ("https://[2001:db8::1]/status", 1000 - 972), + ("https://[2001:db8:85a3::8a2e:370:7334]:443/path?ok=1", 1000 - 948), + ("mailto:someone@example.com", 1000 - 974), + ("git+ssh://git@example.co.uk:22/repo.git", 1000 - 961), + ("https://very.long.tld.example.museum/collection/item", 1000 - 977), + + # Usernames - local and remote + ("@alice", 6), + ("@bob@example.com", 4), + ("hi @charlie and @dora@example.social!", 1000 -978), + + # Mixed + ("hey @me@example.com look at https://example.com/a-b_c~d?e=f#g and @you ", 50), + + # Grapheme cluster vs code point differences + ("a: 🇪🇪", 4), + ("b: 👨‍👩‍👧‍👦", 4), + ("c: 👩🏽‍💻", 4), + ("d: ✊🏿", 4), + ("é", 1), + ("f\u0301", 1), + + # Stress-tests + ("https://sub.sub2.пример.рф/путь/страница?параметр=значение#якорь", 47), + ("clusters: 😀😃😄😁😆😅😂🤣😊🙂😉🙃😇🥰😍🤩😘😗😙😚", 30), + + # Varied compositions + ("See: http://example.com https://[2001:db8::2]:8443/a ftp://user:pw@files.example.org:21/x http://192.168.0.1/", 1000 - 886), + ("@one https://example.social/@two 👩🏽‍💻 🇪🇪 @three@example.com ✊🏿", 1000 - 959), + + # Edge punctuation around URLs/usernames + ("(see https://example.com.)", 30), + ("[link: http://user:pass@host.example:8080/path?x=y#z]", 1000 - 947), + ("<@root> and {@admin@example.net}", 20), + ("https://example.com/a-b_c~d?param_a=1¶m-b=2", 1000 - 977), +] + +@pytest.mark.parametrize("text,expected", TEST_CASES) +def test_get_status_length_against_ground_truth(text, expected): + assert Mastodon.get_status_length(text) == expected + assert Mastodon.get_status_length(text, "what") == expected + 4 diff --git a/tox.ini b/tox.ini index c7f40dc..f31abec 100644 --- a/tox.ini +++ b/tox.ini @@ -4,5 +4,5 @@ skipsdist = true [testenv] -deps = .[test,webpush,blurhash] +deps = .[test,webpush,blurhash,grapheme] commands = python setup.py test