Allow returning page if transcript extraction failed
| @@ -22,59 +22,15 @@ const visitHeaders = { | |||
| 22 | 'Sec-Fetch-User': '?1', | 22 | 'Sec-Fetch-User': '?1', |
| 23 | }; | 23 | }; |
| 24 | 24 | ||
| 25 | router.post('/serpapi', jsonParser, async (request, response) => { | ||
| 26 | try { | ||
| 27 | const key = readSecret(request.user.directories, SECRET_KEYS.SERPAPI); | ||
| 28 | |||
| 29 | if (!key) { | ||
| 30 | console.log('No SerpApi key found'); | ||
| 31 | return response.sendStatus(400); | ||
| 32 | } | ||
| 33 | |||
| 34 | const { query } = request.body; | ||
| 35 | const result = await fetch(`https://serpapi.com/search.json?q=${encodeURIComponent(query)}&api_key=${key}`); | ||
| 36 | |||
| 37 | console.log('SerpApi query', query); | ||
| 38 | |||
| 39 | if (!result.ok) { | ||
| 40 | const text = await result.text(); | ||
| 41 | console.log('SerpApi request failed', result.statusText, text); | ||
| 42 | return response.status(500).send(text); | ||
| 43 | } | ||
| 44 | |||
| 45 | const data = await result.json(); | ||
| 46 | return response.json(data); | ||
| 47 | } catch (error) { | ||
| 48 | console.log(error); | ||
| 49 | return response.sendStatus(500); | ||
| 50 | } | ||
| 51 | }); | ||
| 52 | |||
| 53 | /** | 25 | /** |
| 54 | * Get the transcript of a YouTube video | 26 | * Extract the transcript of a YouTube video |
| 55 | * @copyright https://github.com/Kakulukian/youtube-transcript (MIT License) | 27 | * @param {string} videoPageBody HTML of the video page |
| 28 | * @param {string} lang Language code | ||
| 29 | * @returns {Promise<string>} Transcript text | ||
| 56 | */ | 30 | */ |
| 57 | router.post('/transcript', jsonParser, async (request, response) => { | 31 | async function extractTranscript(videoPageBody, lang) { |
| 58 | try { | ||
| 59 | const he = require('he'); | 32 | const he = require('he'); |
| 60 | const RE_XML_TRANSCRIPT = /<text start="([^"]*)" dur="([^"]*)">([^<]*)<\/text>/g; | 33 | const RE_XML_TRANSCRIPT = /<text start="([^"]*)" dur="([^"]*)">([^<]*)<\/text>/g; |
| 61 | const id = request.body.id; | ||
| 62 | const lang = request.body.lang; | ||
| 63 | const json = request.body.json; | ||
| 64 | |||
| 65 | if (!id) { | ||
| 66 | console.log('Id is required for /transcript'); | ||
| 67 | return response.sendStatus(400); | ||
| 68 | } | ||
| 69 | |||
| 70 | const videoPageResponse = await fetch(`https://www.youtube.com/watch?v=${id}`, { | ||
| 71 | headers: { | ||
| 72 | ...(lang && { 'Accept-Language': lang }), | ||
| 73 | 'User-Agent': visitHeaders['User-Agent'], | ||
| 74 | }, | ||
| 75 | }); | ||
| 76 | |||
| 77 | const videoPageBody = await videoPageResponse.text(); | ||
| 78 | const splittedHTML = videoPageBody.split('"captions":'); | 34 | const splittedHTML = videoPageBody.split('"captions":'); |
| 79 | 35 | ||
| 80 | if (splittedHTML.length <= 1) { | 36 | if (splittedHTML.length <= 1) { |
| @@ -129,11 +85,73 @@ router.post('/transcript', jsonParser, async (request, response) => { | |||
| 129 | })); | 85 | })); |
| 130 | // The text is double-encoded | 86 | // The text is double-encoded |
| 131 | const transcriptText = transcript.map((line) => he.decode(he.decode(line.text))).join(' '); | 87 | const transcriptText = transcript.map((line) => he.decode(he.decode(line.text))).join(' '); |
| 88 | return transcriptText; | ||
| 89 | } | ||
| 90 | |||
| 91 | router.post('/serpapi', jsonParser, async (request, response) => { | ||
| 92 | try { | ||
| 93 | const key = readSecret(request.user.directories, SECRET_KEYS.SERPAPI); | ||
| 94 | |||
| 95 | if (!key) { | ||
| 96 | console.log('No SerpApi key found'); | ||
| 97 | return response.sendStatus(400); | ||
| 98 | } | ||
| 99 | |||
| 100 | const { query } = request.body; | ||
| 101 | const result = await fetch(`https://serpapi.com/search.json?q=${encodeURIComponent(query)}&api_key=${key}`); | ||
| 102 | |||
| 103 | console.log('SerpApi query', query); | ||
| 132 | 104 | ||
| 105 | if (!result.ok) { | ||
| 106 | const text = await result.text(); | ||
| 107 | console.log('SerpApi request failed', result.statusText, text); | ||
| 108 | return response.status(500).send(text); | ||
| 109 | } | ||
| 110 | |||
| 111 | const data = await result.json(); | ||
| 112 | return response.json(data); | ||
| 113 | } catch (error) { | ||
| 114 | console.log(error); | ||
| 115 | return response.sendStatus(500); | ||
| 116 | } | ||
| 117 | }); | ||
| 118 | |||
| 119 | /** | ||
| 120 | * Get the transcript of a YouTube video | ||
| 121 | * @copyright https://github.com/Kakulukian/youtube-transcript (MIT License) | ||
| 122 | */ | ||
| 123 | router.post('/transcript', jsonParser, async (request, response) => { | ||
| 124 | try { | ||
| 125 | const id = request.body.id; | ||
| 126 | const lang = request.body.lang; | ||
| 127 | const json = request.body.json; | ||
| 128 | |||
| 129 | if (!id) { | ||
| 130 | console.log('Id is required for /transcript'); | ||
| 131 | return response.sendStatus(400); | ||
| 132 | } | ||
| 133 | |||
| 134 | const videoPageResponse = await fetch(`https://www.youtube.com/watch?v=${id}`, { | ||
| 135 | headers: { | ||
| 136 | ...(lang && { 'Accept-Language': lang }), | ||
| 137 | 'User-Agent': visitHeaders['User-Agent'], | ||
| 138 | }, | ||
| 139 | }); | ||
| 140 | |||
| 141 | const videoPageBody = await videoPageResponse.text(); | ||
| 142 | |||
| 143 | try { | ||
| 144 | const transcriptText = await extractTranscript(videoPageBody, lang); | ||
| 133 | return json | 145 | return json |
| 134 | ? response.json({ transcript: transcriptText, html: videoPageBody }) | 146 | ? response.json({ transcript: transcriptText, html: videoPageBody }) |
| 135 | : response.send(transcriptText); | 147 | : response.send(transcriptText); |
| 136 | } catch (error) { | 148 | } catch (error) { |
| 149 | if (json) { | ||
| 150 | return response.json({ html: videoPageBody, transcript: '' }); | ||
| 151 | } | ||
| 152 | throw error; | ||
| 153 | } | ||
| 154 | } catch (error) { | ||
| 137 | console.log(error); | 155 | console.log(error); |
| 138 | return response.sendStatus(500); | 156 | return response.sendStatus(500); |
| 139 | } | 157 | } |