mirror of
https://gitlab.rlp.net/proj-wise2526-video2document/video2document.git
synced 2026-06-15 18:01:52 +02:00
Summarizer und weiters angepasst
This commit is contained in:
@@ -7,6 +7,7 @@ if (!fs.existsSync(outputDir)) {
|
|||||||
fs.mkdirSync(outputDir, { recursive: true });
|
fs.mkdirSync(outputDir, { recursive: true });
|
||||||
}
|
}
|
||||||
|
|
||||||
|
//Speaker, ALL-Sentences, Start, End
|
||||||
|
|
||||||
module.exports = {
|
module.exports = {
|
||||||
name: "summarize-transcription", // Unique name for our function that will later be used to get the function from the map via "mapFunctions.get("example").function()"
|
name: "summarize-transcription", // Unique name for our function that will later be used to get the function from the map via "mapFunctions.get("example").function()"
|
||||||
@@ -34,8 +35,7 @@ module.exports = {
|
|||||||
return { error: "Invalid JSON" };
|
return { error: "Invalid JSON" };
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
// console.log(inputJson);
|
|
||||||
console.log(outputDir);
|
|
||||||
const words = inputJson.words;
|
const words = inputJson.words;
|
||||||
if (!Array.isArray(words)) {
|
if (!Array.isArray(words)) {
|
||||||
return { error: "No words Array found" };
|
return { error: "No words Array found" };
|
||||||
|
|||||||
@@ -0,0 +1,121 @@
|
|||||||
|
const fs = require("fs");
|
||||||
|
const path = require("path");
|
||||||
|
|
||||||
|
// Prepare output directory (always storage/transcriptionSummaries under project root)
|
||||||
|
const outputDir = `${__dirname}/../../../storage/transcriptionSummaries`;
|
||||||
|
if (!fs.existsSync(outputDir)) {
|
||||||
|
fs.mkdirSync(outputDir, { recursive: true });
|
||||||
|
}
|
||||||
|
|
||||||
|
//Speaker, Sentence, Start, End
|
||||||
|
|
||||||
|
module.exports = {
|
||||||
|
name: "summarize-transcription2", // Unique name for our function that will later be used to get the function from the map via "mapFunctions.get("example").function()"
|
||||||
|
type: "summarizer", // value used to differentiate each module to order them in the UI
|
||||||
|
displayname: "Summarizer", // The displayname used within the UI
|
||||||
|
async function(args) {
|
||||||
|
let inputJson = args.json;
|
||||||
|
|
||||||
|
//JSON Path
|
||||||
|
if (args.jsonPath) {
|
||||||
|
try {
|
||||||
|
const raw = fs.readFileSync(args.jsonPath, "utf-8");
|
||||||
|
inputJson = JSON.parse(raw);
|
||||||
|
} catch (e) {
|
||||||
|
console.error("Failed to load JSON from file:", e);
|
||||||
|
return { error: "Could not read JSON from file path." };
|
||||||
|
}
|
||||||
|
}
|
||||||
|
// JSON parsen
|
||||||
|
if (typeof inputJson === "string") {
|
||||||
|
try {
|
||||||
|
inputJson = JSON.parse(inputJson);
|
||||||
|
} catch (e) {
|
||||||
|
console.log("Invalid JSON in summarize-transcription");
|
||||||
|
return { error: "Invalid JSON" };
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
const words = inputJson.words;
|
||||||
|
if (!Array.isArray(words)) {
|
||||||
|
return { error: "No words Array found" };
|
||||||
|
}
|
||||||
|
|
||||||
|
const ENDINGS = [".", "!", "?"]; // '...' auch als Satzende ?
|
||||||
|
const ABBREVIATIONS = new Set(["z.B.", "bzw.", "u.a.", "Dr.", "Mr.", "Mrs.", "Prof.", "etc."]); //TODO weitere Ergaenzen
|
||||||
|
|
||||||
|
const result = [];
|
||||||
|
let currentSentence = "";
|
||||||
|
let currentSpeaker = null;
|
||||||
|
let startTime = null;
|
||||||
|
let endTime = null;
|
||||||
|
|
||||||
|
for (const w of words) {
|
||||||
|
if (!currentSpeaker) currentSpeaker = w.speaker;
|
||||||
|
if (startTime === null) startTime = w.start;
|
||||||
|
endTime = w.end;
|
||||||
|
|
||||||
|
//speaker changing
|
||||||
|
if (currentSpeaker !== w.speaker && currentSentence) {
|
||||||
|
result.push({
|
||||||
|
speaker: currentSpeaker,
|
||||||
|
sentence: currentSentence,
|
||||||
|
start: startTime,
|
||||||
|
end: endTime
|
||||||
|
});
|
||||||
|
currentSentence = "";
|
||||||
|
startTime = w.start;
|
||||||
|
}
|
||||||
|
currentSpeaker = w.speaker;
|
||||||
|
currentSentence += (currentSentence ? " " : "") + w.text; //sentence beginning or not
|
||||||
|
const lastWord = w.text.trim();
|
||||||
|
const lastChar = lastWord.slice(-1);
|
||||||
|
const isAbbreviation = ABBREVIATIONS.has(lastWord);
|
||||||
|
|
||||||
|
//sentence ending
|
||||||
|
if (ENDINGS.includes(lastChar) && !isAbbreviation) {
|
||||||
|
result.push({
|
||||||
|
speaker: currentSpeaker,
|
||||||
|
sentence: currentSentence,
|
||||||
|
start: startTime,
|
||||||
|
end: endTime
|
||||||
|
});
|
||||||
|
currentSentence = "";
|
||||||
|
startTime = null;
|
||||||
|
endTime = null;
|
||||||
|
currentSpeaker = null;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
// safe last sentence
|
||||||
|
if (currentSentence) {
|
||||||
|
result.push({
|
||||||
|
speaker: currentSpeaker,
|
||||||
|
sentence: currentSentence,
|
||||||
|
start: startTime,
|
||||||
|
end: endTime
|
||||||
|
});
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
// Output as Text
|
||||||
|
const output = result.map(r =>
|
||||||
|
`Sprecher ${r.speaker} [${r.start.toFixed(2)} - ${r.end.toFixed(2)}]: ${r.sentence}`
|
||||||
|
);
|
||||||
|
|
||||||
|
// Output on cosole
|
||||||
|
//console.log("\n------------\nMerged Transcription Result:\n", output, "\n------------\n");
|
||||||
|
|
||||||
|
try {
|
||||||
|
const jsonPath = path.join(outputDir, "transcription_result.json");
|
||||||
|
fs.writeFileSync(jsonPath, JSON.stringify(result, null, 2), "utf-8");
|
||||||
|
|
||||||
|
const txtPath = path.join(outputDir, "transcription_result.txt");
|
||||||
|
fs.writeFileSync(txtPath, output.join("\n"), "utf-8");
|
||||||
|
|
||||||
|
console.log(`Summary successfully saved:\n- ${jsonPath}\n- ${txtPath}`);
|
||||||
|
} catch (err) {
|
||||||
|
console.error("Error saving Summary:", err);
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
@@ -1,9 +1,6 @@
|
|||||||
//require('dotenv').config();
|
require('dotenv').config();
|
||||||
//const axios = require('axios');
|
|
||||||
//const fs = require('fs');
|
|
||||||
const path = require('path');
|
|
||||||
|
|
||||||
const API_KEY = process.env.ASSEMBLYAI_API_KEY;
|
const API_KEY = process.env.API_KEY;
|
||||||
const BASE_URL = 'https://api.assemblyai.com/v2';
|
const BASE_URL = 'https://api.assemblyai.com/v2';
|
||||||
|
|
||||||
//---------------------------------------------------Upload audio---------------------------------------------------
|
//---------------------------------------------------Upload audio---------------------------------------------------
|
||||||
@@ -93,7 +90,7 @@ module.exports = {
|
|||||||
type: 'transcription',
|
type: 'transcription',
|
||||||
displayname: 'AssemblyAI',
|
displayname: 'AssemblyAI',
|
||||||
|
|
||||||
run: async (audioFileName) => {
|
async function(audioFileName) {
|
||||||
try {
|
try {
|
||||||
// audioFileName ist nur "datei.mp3"
|
// audioFileName ist nur "datei.mp3"
|
||||||
const audioPath = path.join(
|
const audioPath = path.join(
|
||||||
|
|||||||
@@ -6,6 +6,13 @@ module.exports = {
|
|||||||
// We are now calling the example function from the example folder
|
// We are now calling the example function from the example folder
|
||||||
mapFunctions.get("example").function("Startup")
|
mapFunctions.get("example").function("Startup")
|
||||||
|
|
||||||
|
let transcript = await mapFunctions.get("assembly").function('../../storage/audio/IMG_2978.wav');
|
||||||
|
|
||||||
|
let summary = await mapFunctions.get("summarize-transcription").function({jsonPath:'/Users/santa/Proj25/video2document/storage/transcripts/IMG_2978.json'});
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./a.mp4", outputType:"wav"})
|
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./a.mp4", outputType:"wav"})
|
||||||
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./b.mp4", outputType:"wav"})
|
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./b.mp4", outputType:"wav"})
|
||||||
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./b.mp4", outputType:"flac"})
|
// mapFunctions.get("extraction-video-to-audio").function({inputVideoPath:"./b.mp4", outputType:"flac"})
|
||||||
|
|||||||
@@ -0,0 +1,18 @@
|
|||||||
|
require('dotenv').config();
|
||||||
|
const path = require('path');
|
||||||
|
const assemblyModule = require('../../services/modules/transcription-remote/assembly.js');
|
||||||
|
|
||||||
|
// Audio-Datei oder URL aus Kommandozeile, Standard: test.wav
|
||||||
|
const audioPath = process.argv[2] || './storage/audio/IMG_2978.wav';
|
||||||
|
|
||||||
|
(async () => {
|
||||||
|
try {
|
||||||
|
const transcript = await assemblyModule.run(audioPath);
|
||||||
|
|
||||||
|
console.log('Transcription succesful');
|
||||||
|
console.log('Transcript ID:', transcript?.id);
|
||||||
|
console.log('Speaker labels:', transcript?.utterances?.length || 0);
|
||||||
|
} catch (error) {
|
||||||
|
console.error('Error in Transcription:', error?.message || error);
|
||||||
|
}
|
||||||
|
})();
|
||||||
Reference in New Issue
Block a user