Replying to Avatar 阿白

npub16dq3qps8sgehezrylamryx5zrukmhnashvecvjcucjr3927cf2qqd8evcw

npub1ly0gm0wep64xngzxv4fup9rhxtrw5l2e87djmct32s8d4eaz3z8qqca5wc

npub19l03suyl2h40ym3zhscrseh346ug3u22ptt88syps27hj9vfqfxq3ykck2

npub10vup35yg4zrtsq0phw5k956sqxs860vw8u0gn52m67apvy72lktq8xj7tf

npub1rw07v4m9q007w5l9uelkgsjtylpwz2csvzhzfp0xv66m3lyeegnq4fpemg

npub1d5ygkef6r0l7w29ek9l9c7hulsvdshms2qh74jp5qpfyad4g6h5s4ap6lz

npub1nr06h4rzntertpj2lzpgje6mxt9hetjk5ffn78suy9525esqn0rq2wnzyk

npub15n4yau07kv2asacnteut7jnzy3ke2cwcc3vcvrzhkwr4pqa3v7xqrngqfv

npub1xv00thmc5qr02pcrmgss8z29dnn2x9yr5nszwf0sjzfqvkvk2l2qkvwsmg

npub1vj0wlergmkcs0sz7hfks2ywj555c2s87f40squ4sqcmqpr7897fqn6mfew

npub12w5rjt5hrdrrym3ap7yk0kchcnmuefx5905hnvtxgyjv3a56755qx6u7fz

npub1pjvvr9we639fn6kp9aqnf4rua5suwe8mdvfygr0gegadhuxrwwnsz5jf67

npub1924489x2wenugrngjwtlded3j29rk2ffr7e4an3v0ufzgt5xw03s3st05g

npub1recxqywfx2aphqqd22rujxjj0urqsghgsk8xdtrkal7e0s20067swludg8

npub1ss2ndfhhsk6frk7sfucqhe7flr6um0dj24yelxlf0h0dhtk8m9qs4t6j68

npub1d80u7884h6qsjrvaj5c5l7upuq3sh86kn56sev46hesg6na27wcq62tflx

npub1k74g03fmrdcxun6x3qvrwq8mpfg40kagnhlez6rh2pft65l8asmqzs83r8

npub1hvpdqkq7ysg5ym4ezmnnsms0w7ak9ptwhsg279y6sz0du8v62ncqx2puhw

npub127fmadjnpq55fhx448ec6sva6ppkc9ghzvasc3zy4km9s5uljagsaathf5

npub19786pedeqk8mpj9favvzxrz44h6d3d3c6ur2fe9592cgqljm9g7qcz8jjs

npub1kuzfar30lcclpq7049elu0s0mln2lwjgl2yjsy0zqj0k3nqvfvnql6ul7r

npub19ztvppq4vt52965efcc6un0spxj8ndtya87e7e3xu2ratdwg547sy5qh4g

npub1tvy9mx0lsl9tlypll3lk0mswxwrfv0zpc78n7vruh4uvm3pylzkq62l74m

npub1ly0gm0wep64xngzxv4fup9rhxtrw5l2e87djmct32s8d4eaz3z8qqca5wc

npub104k6nlkgugz98j92ra2246387c6swln5dgex4ur203vltg9jjaqs6c76vv

npub1724ns8dm5e5jdyas44zvajyvy8dtdqvqwy69xvcamuld3nflve0qcm87gn

npub1a9h5u0td2u059e2thq5t9xlc9xhlqaw38c8levd29maaaqul92tqkkuyxh

npub1vugxyx27dtjxg0vj6ydehwkfwujg749vzfn9hwsyrm78pc65c72s3m7mlc

npub1yd9syffrjrfmdmu2kj4upa4jdvuqhcrreqdwtyc6pf8380gwanksauar9d

npub1fanaslw4azll8dqked96q8k9m8gu3ktm74enxnfq2esp5dd7unvqrq7r2l

npub1x8dzy9xegwmdk2vy30l8u08caspcqq2yzncxehdsa6kvnte9pr3qnt8pg4

npub125xsgn2hvp3rkzv3n0u04hwqfuq4r3wl52vngvjs4y3kf59zr36qex8k6j

npub1tuqsl6l8xzly95vv80um7wsnt7gxy8w9wgt4khp4wyv4xwhfw44slm93e9

npub1h3nwpm6ks4kz8vydws5tug2sgju99xmzpqkgx3fce03ca99s62kqh98ajl

npub1448pmcdknujz2wyqrua4fehdhaw8vc8vwgfz6dvrzuuukvj8gs4s29ns27

npub1zj6qapky0ydurlrd75nkwstymhsl9g520ddhkfgu8h5fdga9nrhsenncg5

npub1ydgm8xkac306vcjxye8sg9znh5lz4jjmgenczwp3rc9rll209kaqewkjwk

npub1cxsf9k3fgm7d9ukwgch8ftj9fgw8w4dy2986xgjhesfkkdywyfeqpz8g5p

npub1sxtdqgnqvwv8w3x8s3490p7s7wd6mvas3sqmunradd4k0mntmcps8xp97m

npub1304fkk8h0lu8cqe539627qedfusejcq6hdrarg4ymtnpfduc00zsgh4qpu

npub1u4qayfnkfd8rtrk9t0turnnk4a8ksr3hcfsxvgjnzdel05d45mdqu33rp3

npub1lsdw425utjhujnkr0nyzg7d7h262wdz6zmmh34d2sxarlkcmyzlsm7cccz

npub15zytpa8xdr9hesxgrjxnypseyc3549jye843ax8vqq5ym9r9l0ysrrfksh

npub1d5yu6h2le9z290528f2z84n7elfe85cm2cjujvjmgkqcjwd36tts229pe5

npub1d37f0nlgn7nyeawkelnxne2see0x52jjfplq4u4n6hwfex5rx24sgdy7xy

npub1k5yxkxvjuz36u8lh3rz7mvu285d9xkqv5y2tk0ewfqpr4mvmshzqdplxnl

npub1y60x74a2ngsqeq2wdwv8yxqemhsr3jnqcqushpaktrfsp2mdp5zqe9cll0

npub1m8jn24mf7vk8yhz86d25yj6gf3pk7nqy490ztxfkywjrnv5uk7rq6hynsv

npub1k72g25az0lrn6dxht0yzju0f69v8h9g3hrdt8yf737wydq5sff0qhdc2k0

npub1d3jpxhjlu2xkv2r9mk6a3vceghk8jex9uh6qdzss30jlvscxkk6skfgwt3

npub18dyedpz80p9lersghgn70jxpm577mwtvk30stzge2q40cfxy2phs5mt94x

npub1et8kzrwzg2czjsrw6m2cy0m3kdefg5amtms50tvsrs2fg9ywc65s9prrva

npub1lsdahrfsj55f3yynj4dyx73g4fmu5hwvepk0rv7zs4w9zsnt4jnqg2xd76

npub1cyfcr2xja8qt40asu4s3u50m3hrg6pf4lf64t242jthqveca65esjpw957

npub1h9sncyp0fqfx4jkucutzuy2rvpmyrerey5x5pj8sysz0pvqm3pmqmvl4ge

npub1egr8gdszm2kg07t8lw5qrmdn9zgsslhwektqga9u3n32ec0v8jysgymhpz

npub1lwdt2m9xp8pcleellcdpkyf973mfwlxjfcdvs47jldatc60c6gfqzlf5rx

npub130kmga94aynxnmlxac9lt0kx5pp67xckt0xelst6xkx0a6rxn0qskrtgce

npub1nh8e740yexfl87cyqx9keqk7hw2e37r9fjsqt6t8sp5rar25e4mq9mu737

npub15lgnqn046mmru6jvsjezzew2pdwh29g88tnt9ah6kmd3js3dq6nqywyq90

npub1gzumvszm3x3ran4euqu3pvu3p4znzry6v55z3apgaxjwhznjyu7svk5cyg

npub1szsdl6ah7km96z28auesrkpgah25ww4xkknert722avrcduh3zas2le827

npub1l5unz0t8gjpzd4j9ckyafk69j4exzva6qx4ee7rtvra5sxhsrg7qg4tpp0

npub1fzzml46kdmlcyxq5n99eap2h24k0x2nmx58dszhwyhrcc5k667jsu46urj

npub1pjd3a8l0wmygkclcvezacvamwamlqfv7cs0xwjmp7n7920mdkrsq9hpeqp

npub14phyutfe0lfzpr6mt2trndzfl278svheheh9tvqqpaynujvm05lqkjw08t

npub1958rytvzewvz0dmk2yjpy0fgeddzagww740fmgj85frqhdwvpzxqmgj6mm

npub1x2hcl8jtnqe9dycuk4240lcc3t088tytl37dm5m2hqslyqewgp0sq6d8v7

npub1x6gqev6j2lqwl6qa6kx9ekq8gkwmyyj3x9ayf9mkc4agrmsg55xsht4wmn

npub10sagg39fsskssvxm57x4pjjtx6t69x6lz0xvxq5nl9pd4e49cd7szqaemm

npub1pl05tgwynp3jk5wyq3l5e8fvu995rzqjmcly32f2ptua8w9rx6zsjuvyfl

npub1qayh77kdlyhzk9vtmadyslefv3yuvcvlfe5gdl2rlw66ljgj63ss7j5h9k

里面有一些是日文用户

Reply to this note

Please Login to reply.

Discussion

# 定义一个函数,检查字符串中是否包含中文字符

def contains_chinese(s):

for ch in s:

if u'\u4e00' <= ch <= u'\u9fff':

return True

return False

这个判断字符串中是否包含中文字符的算法,会把日文看成中文😂

因为日语中会有纯汉字,编码是一模一样的。

所以应该还要结合profile和其它的post来筛选,出现过片假名就可以一票否决了。

这个是一个日本用户的元数据,是否可以用name来判断是否包含片假名,来检测是否是日本用户?

{"id":"1ede94a90be721b027eac95b6671547d2938b0b3a57bbc36c405d829766c41d3","kind":0,"pubkey":"269e6f57aa9a200c814e6b98721819dde038ca60c0390b87b658d300ab6d0d04","tags":[],"created_at":1679053245,"content":"{\"name\":\"ゆうひまわり\",\"nip05\":\"yuuhimawari@yuuhimawariero.net\",\"picture\":\"\",\"banner\":\"\",\"about\":\"illust/イラスト/格ゲー\",\"lud06\":\"\",\"lud16\":\"itchyfood23@walletofsatoshi.com\",\"display_name\":\"ゆうひまわり\",\"nip05valid\":true,\"webOfTrustScore\":11.799999999999974,\"followingCount\":179,\"followersCount\":135}","sig":"fe55e192413ff26d94ed966564a6a0278f15b7e5414f06038b7440e28ec650574006573c90e6e95e78cb4f4e5923c3c077a61972389f993eebbf35caff41124b"}

name displayname about 过往的各种post,任何地方检测到比如超过30%的假名就可以否决了,可以试试不同阈值。

综合考量的信息越多越准确,其实最好的方法还是基于贝叶斯概率分类,你手动标注几个之后,就能算出来一个用户是中文用户还是日文用户的概率,然后你给定一个阈值即可,比如95%以上概率是中文用户的收集出来。

这事儿@npub1vj0wlergmkcs0sz7hfks2ywj555c2s87f40squ4sqcmqpr7897fqn6mfew 熟悉啊,他的relay支持语言过滤功能。