Merge branch 'develop' into 'feature/meeting_document_types'

# Conflicts:
#   electron/main/script.js
#   main.js

.gitlab-ci.yml

@@ -24,5 +24,4 @@ job-test:
     - npm install
     - echo "ASSEMBLYAI_API_KEY=$apikey_assembly" > .env
     - echo "GOOGLE_API_KEY=$apikey_gemini" >> .env
-    - echo "SAIA_API_KEY=$apikey_saia" >> .env
     - npm test

electron/main/index.html

@@ -19,8 +19,8 @@
       </label>
 
       <nav class="menu1">
-        <a href="index2.html" class="li1">Custom document</a>
-        <a href="index3.html" class="li1">Help</a>
+        <li class="li1">Help</li>
+        <li class="li1">Language</li>
       </nav>
     </nav>
   </section>
@@ -79,7 +79,7 @@
       </div>
     </div>
 
-    <div class="step" id="step3" style="display:none;">
+    <div class="step" id="step3" style="display:none;"> 
       <div class="checkbox-group">
         <label id="checkbox-label" for="checkbox-group">Choose prefered document style:</label>
         <div class="checkbox-container">

electron/main/script.js

@@ -27,12 +27,22 @@ function checkBoxes() {
       if(testEndings.some(e => pathToLower.endsWith(e))){
         //assembly of the json for the main
 
-        var typeCheckbox;
-        if(document.getElementById("docFormat").checked) typeCheckbox = document.getElementById("docFormat").value;
-        if(document.getElementById("docFormatSummary1").checked) typeCheckbox = document.getElementById("docFormatSummary1").value;
-        if(document.getElementById("docFormatSummary2").checked) typeCheckbox = document.getElementById("docFormatSummary2").value;
-        if(document.getElementById("docFormatSummary3").checked) typeCheckbox = document.getElementById("docFormatSummary3").value;
-        if(document.getElementById("docFormatCustom").checked) typeCheckbox = document.getElementById("docFormatCustom").value;
+        /*
+        const selectedStyles = [checkedCounter];
+        var iter = 0;
+        checkboxes.forEach(function(checkbox){
+        if(checkbox.checked){
+          console.log(checkbox.value);
+          selectedStyles[iter] = {iter: checkbox.value};
+          iter++;
+          } 
+        });
+        */
+ 
+        function getSelectedDocumentType() {
+          const checked = document.querySelector('input[name="docFormat"]:checked');
+          return checked ? checked.value : null;
+        }
 
         document.getElementById("testy").style.visibility = "visible"
         document.getElementById("box1").style.backgroundColor = "red";
@@ -53,8 +63,8 @@ function checkBoxes() {
           },
           "document": {
             "module":aiType.value,
-            "type": typeCheckbox,
-            "outputType": outputType.value
+            "type": getSelectedDocumentType()
+            //"styles": selectedStyles
           }
         };
         window.submit.submit(sendingPackage)
@@ -302,7 +312,7 @@ function setSpeakerAudiosValue(valy){
   try {
     speakerAudios = valy;
     speakerRewriten = valy;
-    document.getElementById("speakerAudioViewer").src = valy.speakerA.src;
+    document.getElementById("speakerAudioViewer").src = valy.speakerA.source;
   } catch (error) {
     
   }

electron/main/style.css

@@ -177,11 +177,10 @@ input[type="file"] {
 .submit-btn {
   display: flex;
   justify-content: center;
-  align-items: center;
   padding: 10px 20px;
-  margin-left: 300px;
+  margin-left: 80px;
   margin-top: 30px;
-  margin-bottom: 70px;
+  margin-bottom: 10px;
   background-color: #007BFF;
   color: white;
   border: none;
@@ -421,6 +420,7 @@ li {
   border-radius: 5px;
   background-color: #1C3B69;
   margin: 0;
+  display: -ms-grid;
   display: grid;
   grid-template-rows: 1fr repeat(4, 0.5fr);
   grid-row-gap: 25px;
@@ -436,21 +436,19 @@ li {
   -webkit-transition: all 0.3s ease;
 }
 
-.menu1 a:first-child {
+.menu1 li:first-child {
   margin-top: 30px;
 }
 
-.menu1 a:last-child {
+.menu1 li:last-child {
   margin-bottom: 30px;
 }
 
 .li1 {
-  color: #fff;
   width: 100%;
   margin: 0;
-  padding: 10px 0px;
+  padding: 10px 0;
   font: 700 20px 'Oswald', sans-serif;
-  text-decoration: none;
 }
 
 .li1:hover {
@@ -469,7 +467,7 @@ li {
   display: flex;
   justify-content: center;
   padding: 10px 20px;
-  margin-left: 10px;
+  margin-left: 80px;
   margin-top: 30px;
   margin-bottom: 10px;
   background-color: #007BFF;

main.js

@@ -194,27 +194,39 @@ electron.ipcMain.on("file_submit", async (event, args) => {
         console.log("\n\n Running the LLM module");
         // TODO implement documentation module
         // This code handles the Text to Document processing module call
-
+        
         console.log(`\n\n Running the LLM for Document Style ${args.document.type}`);
 
         await mapFunctions.get("module-handler").function(args.document.module, { inputTranscriptPath: transcriptpath, documentTypePath: "./storage/documentType/" + templateFile, language: "en" }).then(resp => {
             console.log(resp);
-            globalFinalHtmlPath = resp
+            transcriptpath = resp
             curstep++
-            mainWindow.webContents.send("progress", {curstep:curstep, totalsteps:totalsteps})
+            mainWindow.webContents.send("progress", { curstep: curstep, totalsteps: totalsteps })
         }).catch(err => {
             mainWindow.webContents.send("error", err)
             return
         })
+        
+
+        // TODO actually implement this functionality
+        // Module to get the first few lines for each speaker to send to the frontend
+//         await mapFunctions.get("speaker-getter-idfk").function(transcriptpath).then(resp => {
+//             console.log(resp);
+//             transcriptpath = resp
+//             curstep++
+//             mainWindow.webContents.send("progress", {curstep:curstep, totalsteps:totalsteps})
+
+// // {
+// // speakerA: {source: "Pfad zur Audio File"},
+// // speakerB:.....
+// // }
+            mainWindow.webContents.send("speakers", {speakerA:"pfad1", speakerB:"pfad2"})
+//         }).catch(err => {
+//             mainWindow.webContents.send("error", err)
+//             return
+//         })
 
 
-        await mapFunctions.get("extract-speaker-snippets").function({audioPath: audiopath, jsonPath: transcriptpath }).then(resp => {
-            mainWindow.webContents.send("submitSpeaker", resp)
-            console.log(resp)
-        }).catch(err => {
-            mainWindow.webContents.send("error", err)
-            return
-        })
     } catch (error) {
         console.log(error);
     }
@@ -255,4 +267,4 @@ let q1 = {
         {name:"abc", displayname:"ABC"},
         {name:"qeg", displayname:"aqghegahu"}
     ]
-}
+}

package-lock.json

@@ -15,7 +15,7 @@
         "dotenv": "^17.2.3",
         "electron": "^39.1.1",
         "express": "^5.1.0",
-        "ffmpeg-static": "^5.3.0",
+        "ffmpeg-static": "^5.2.0",
         "fluent-ffmpeg": "^2.1.3",
         "html-to-docx": "^1.8.0",
         "mocha": "^11.7.5",
@@ -450,6 +450,7 @@
       "resolved": "https://registry.npmjs.org/@types/node/-/node-24.9.2.tgz",
       "integrity": "sha512-uWN8YqxXxqFMX2RqGOrumsKeti4LlmIMIyV0lgut4jx7KQBcBiW6vkDtIBvHnHIquwNfJhk8v2OtmO8zXWHfPA==",
       "license": "MIT",
+      "peer": true,
       "dependencies": {
         "undici-types": "~7.16.0"
       }
@@ -1305,7 +1306,8 @@
       "version": "0.0.1534754",
       "resolved": "https://registry.npmjs.org/devtools-protocol/-/devtools-protocol-0.0.1534754.tgz",
       "integrity": "sha512-26T91cV5dbOYnXdJi5qQHoTtUoNEqwkHcAyu/IKtjIAxiEqPMrDiRkDOPWVsGfNZGmlQVHQbZRSjD8sxagWVsQ==",
-      "license": "BSD-3-Clause"
+      "license": "BSD-3-Clause",
+      "peer": true
     },
     "node_modules/diff": {
       "version": "4.0.2",
@@ -1771,9 +1773,9 @@
       }
     },
     "node_modules/ffmpeg-static": {
-      "version": "5.3.0",
-      "resolved": "https://registry.npmjs.org/ffmpeg-static/-/ffmpeg-static-5.3.0.tgz",
-      "integrity": "sha512-H+K6sW6TiIX6VGend0KQwthe+kaceeH/luE8dIZyOP35ik7ahYojDuqlTV1bOrtEwl01sy2HFNGQfi5IDJvotg==",
+      "version": "5.2.0",
+      "resolved": "https://registry.npmjs.org/ffmpeg-static/-/ffmpeg-static-5.2.0.tgz",
+      "integrity": "sha512-WrM7kLW+do9HLr+H6tk7LzQ7kPqbAgLjdzNE32+u3Ff11gXt9Kkkd2nusGFrlWMIe+XaA97t+I8JS7sZIrvRgA==",
       "hasInstallScript": true,
       "license": "GPL-3.0-or-later",
       "dependencies": {
@@ -1830,7 +1832,6 @@
       "resolved": "https://registry.npmjs.org/fluent-ffmpeg/-/fluent-ffmpeg-2.1.3.tgz",
       "integrity": "sha512-Be3narBNt2s6bsaqP6Jzq91heDgOEaDCJAXcE3qcma/EJBSy5FB4cvO31XBInuAuKBx8Kptf8dkhjK0IOru39Q==",
       "deprecated": "Package no longer supported. Contact Support at https://www.npmjs.com/support for more info.",
-      "license": "MIT",
       "dependencies": {
         "async": "^0.2.9",
         "which": "^1.1.1"
@@ -4204,6 +4205,7 @@
       "integrity": "sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==",
       "devOptional": true,
       "license": "Apache-2.0",
+      "peer": true,
       "bin": {
         "tsc": "bin/tsc",
         "tsserver": "bin/tsserver"

package.json

@@ -6,7 +6,7 @@
     "dotenv": "^17.2.3",
     "electron": "^39.1.1",
     "express": "^5.1.0",
-    "ffmpeg-static": "^5.3.0",
+    "ffmpeg-static": "^5.2.0",
     "fluent-ffmpeg": "^2.1.3",
     "html-to-docx": "^1.8.0",
     "mocha": "^11.7.5",

services/modules/audioSnippets/extract-speaker-snippets.js

@@ -1,102 +0,0 @@
-
-const ffmpeg = require("fluent-ffmpeg");
-const ffmpegPath = require("ffmpeg-static");
-
-ffmpeg.setFfmpegPath(ffmpegPath);
-
-module.exports = {
-    name: "extract-speaker-snippets",
-    type: "audio",
-    displayname: "Extract Speaker Snippets",
-
-    async function(parameter) {
-        return new Promise(async (resolve, reject) => {
-
-            let output = {}
-
-            // console.log("Extract Speaker Snippets\n");
-
-            // Pfade
-            const AUDIO_PATH = parameter.audioPath; // Gesamt-Audio
-            const JSON_PATH = parameter.jsonPath; // json summary
-            const OUTPUT_DIR = path.join(__dirname, "/../../../storage/audio/speakerSnippets");
-
-
-            if (!AUDIO_PATH || !JSON_PATH) {
-                // console.error("no audioPath or jsonPath available");
-                reject(new Error("no audioPath or jsonPath available"));
-                return;
-            }
-
-            // Output-Ordner
-            if (!fs.existsSync(OUTPUT_DIR)) {
-                fs.mkdirSync(OUTPUT_DIR, { recursive: true });
-            }
-
-            // JSON laden
-            let entries;
-                try {
-                    entries = JSON.parse(fs.readFileSync(JSON_PATH, "utf8"));
-                } catch (err) {
-                    // console.error("JSON reading failed", err);
-                    reject(new Error(err));
-                    return;
-                }
-
-                if (!Array.isArray(entries)) {
-                    // console.error("JSON is not an Array");
-                    reject(new Error("JSON is not an Array"));
-                    return;
-                }
-
-            // Pro Speaker genau EINEN Satz merken
-            const speakerMap = {};
-
-            for (const item of entries) {
-                if (!speakerMap[item.speaker]) {
-                    speakerMap[item.speaker] = item;
-                }
-            }
-
-            // FFmpeg pro Speaker ausführen (sequenziell)
-            for (const speaker of Object.keys(speakerMap)) {
-                const data = speakerMap[speaker];
-
-                // ms → Sekunden
-                const startSec = data.start / 1000;
-                const durationSec = (data.end - data.start) / 1000;
-
-                if (durationSec <= 0) {
-                    // console.log(`invalid times for Speaker ${speaker}`);
-                    continue;
-                }
-
-                const outFile = path.join(OUTPUT_DIR, `speaker_${speaker}.wav`);
-                try {
-                    await new Promise((res, rej) => {
-                        ffmpeg(AUDIO_PATH)
-                            .setStartTime(startSec)
-                            .setDuration(durationSec)
-                            .output(outFile)
-                            .on("end", () => {
-                                output[`speaker${speaker}`] = {src: outFile, name: `speaker${speaker}`}
-                                // console.log(`Snippet erstellt: speaker_${speaker}.wav`);
-                                res();
-                            })
-                            .on("error", (err) => {
-                                // console.error(`FFmpeg Fehler (${speaker})`, err.message);
-                                rej(err);
-                                return
-                            })
-                            .run();
-                    });
-                } catch (error) {
-                    reject(error)
-                    return
-                }
-            }
-            resolve(output)
-            // console.log("\nAlle Speaker-Snippets erstellt\n");
-        })
-    }
-};

services/modules/llm-chat_gpt/chatgpt.js

@@ -1,5 +1,5 @@
-// const fs = require('fs');
-// const path = require('path');
+const fs = require('fs');
+const path = require('path');
 
 const outputDir = path.join(__dirname, "../../../storage/documents"); // path for output directory
 
@@ -43,7 +43,6 @@ const module_exports = {
                 const documentType = await fs.promises.readFile(documentTypePath, "utf-8"); //read document type from Path
                 const promptText = `${documentType}, in language ${language}, transcript:\n\n${transcript}`; //combine doc type, language and transcript - Change prompt here if needed
 
-                // return
                 // --- REST CALL ---
                 const response = await fetch(SAIA_URL, { //safe model response in variable
                     method: "POST",

services/modules/utility/@startup.js

@@ -10,7 +10,6 @@ module.exports = {
     //    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'});
-   //    let snippets = await mapFunctions.get("extract-speaker-snippets").function({audioPath:'/Users/santa/Proj25/video2document/storage/audio/KittyKat.wav', jsonPath1: '/Users/santa/Proj25/video2document/storage/transcriptionSummaries/KittyKat-1765806474958.json' });
 
 
         

storage/documentType/agenda.txt

@@ -21,6 +21,5 @@ STIL:
 - Klar, kompakt
 - Business-orientiert
 - Keine Sprecher- oder Zeitangaben
-- Namen aus dem Transkript speakerA, speakerB etc. sollen weiterhin bestehen bleiben wie sie sind und nicht im Dokument ersetzt werden 
 
 TRANSKRIPT:

storage/documentType/custom_document.txt

@@ -10,7 +10,6 @@ WICHTIG:
 - Nutze das Transkript als Wissensquelle
 - Struktur, Tonalität und Detailgrad anpassen
 - Inhalte logisch zusammenführen
-- Namen aus dem Transkript speakerA, speakerB etc. sollen weiterhin bestehen bleiben wie sie sind und nicht im Dokument ersetzt werden 
 
 FORMAT:
 - Passe Struktur und Stil an den Nutzerwunsch an

storage/documentType/followup_report.txt

@@ -11,7 +11,6 @@ ANFORDERUNGEN:
 - Keine Zeitstempel oder Sprecher-Namen zitieren
 - Leite Entscheidungen und Aufgaben logisch ab, wenn sie implizit sind
 - Markiere offene Punkte klar
-- Namen aus dem Transkript speakerA, speakerB etc. sollen weiterhin bestehen bleiben wie sie sind und nicht im Dokument ersetzt werden
 
 STRUKTUR DES DOKUMENTS:
 1. Titel & Metadaten

storage/documentType/result_protocol.txt

@@ -21,7 +21,5 @@ REGELN:
 - Keine Meinungen oder Spekulationen
 - Keine Zeit- oder Sprecherangaben
 - Sachlich, formal
-- Namen aus dem Transkript speakerA, speakerB etc. sollen weiterhin bestehen bleiben wie sie sind und nicht im Dokument ersetzt werden 
-
 
 TRANSKRIPT:

storage/documentType/sprint_planning_note.txt

@@ -30,6 +30,5 @@ STIL:
 - Agile-konform
 - Klar & umsetzungsorientiert
 - Bullet Points bevorzugen
-- Namen aus dem Transkript speakerA, speakerB etc. sollen weiterhin bestehen bleiben wie sie sind und nicht im Dokument ersetzt werden 
 
 TRANSKRIPT:

test/unit/test.js

@@ -21,7 +21,6 @@ let audiopath
 let transcriptPath
 let summarizePath
 let llmpath
-let speakers
 
 describe("Unit Tests", function() {
 
@@ -35,7 +34,7 @@ describe("Unit Tests", function() {
                 // console.log(resp);
                 done()
             }).catch(err => {
-                done(err);
+                throw err;
             })
         })
         it('Extract .mp4 to .flac', function (done) {
@@ -44,7 +43,7 @@ describe("Unit Tests", function() {
                 // console.log(resp);
                 done()
             }).catch(err => {
-                done(err);
+                throw err;
             })
         })
         it('Extracting to a nonexistant format', function (done) {
@@ -87,7 +86,7 @@ describe("Unit Tests", function() {
                 transcriptPath = resp
                 done()
             }).catch(err => {
-                done(err)
+                throw err
             })
         })
 
@@ -112,7 +111,7 @@ describe("Unit Tests", function() {
             mapFunctions.get("summarize-transcription").function(transcriptPath).then(resp => {
                 done()
             }).catch(err => {
-                done(err)
+                throw err
             })
         })
 
@@ -129,7 +128,7 @@ describe("Unit Tests", function() {
                 summarizePath = resp
                 done()
             }).catch(err => {
-                done(err)
+                throw err
             })
         })
 
@@ -146,92 +145,20 @@ describe("Unit Tests", function() {
         this.slow(30000)
         this.timeout(120000)
 
-        it("ChatGPT", function (done){
-            mapFunctions.get("llm-saia_openai_gpt").function({inputTranscriptPath: summarizePath, documentTypePath: "./storage/documentType/followup_report.txt", language: "en"}).then(resp => {
+        // it("ChatGPT", function (done){
+        //     mapFunctions.get("chatgpt").function({inputTranscriptPath: summarizePath, documentTypePath: "./storage/documentType/meetingReport.json", language: "en"}).then(resp => {
+        //         done()
+        //     }).catch(err => {
+        //         throw err
+        //     })
+        // })
+
+        it("Gemini", function (done){
+            mapFunctions.get("llm-gemini").function({inputTranscriptPath: summarizePath, documentTypePath: "./storage/documentType/meetingReport.json", language: "en"}).then(resp => {
                 llmpath = resp
                 done()
             }).catch(err => {
-                done(err)
-            })
-        })
-
-
-        it("Gemini", function (done){
-            mapFunctions.get("llm-gemini").function({inputTranscriptPath: summarizePath, documentTypePath: "./storage/documentType/followup_report.txt", language: "en"}).then(resp => {
-                done()
-            }).catch(err => {
-                if(err.includes("(503)")){done()} // Error 503 is gemini overload, so an Error that they can at any time throw at us which would crash the pipeline, so we just ignore it and we just imagine that the test passed
-                else{
-                    console.log(err);
-                    done(err)
-                }
-            })
-        })
-
-        it("Qwen3", function (done){
-            mapFunctions.get("qwen3-235b-a22b").function({inputTranscriptPath: summarizePath, documentTypePath: "./storage/documentType/followup_report.txt", language: "en"}).then(resp => {
-                done()
-            }).catch(err => {
-                done(err)
-            })
-        })
-
-
-        it("ChatGPT (Nonexistant Type File)", function (done){
-            mapFunctions.get("llm-saia_openai_gpt").function({inputTranscriptPath: summarizePath, documentTypePath: "a", language: "en"}).then(resp => {
-                done("Didnt crash")
-            }).catch(err => {
-                done()
-            })
-        })
-        it("Gemini (Nonexistant Type File)", function (done){
-            mapFunctions.get("llm-gemini").function({inputTranscriptPath: summarizePath, documentTypePath: "a", language: "en"}).then(resp => {
-                done("Didnt crash")
-            }).catch(err => {
-                done()
-            })
-        })
-        it("Qwen3 (Nonexistant Type File)", function (done){
-            mapFunctions.get("qwen3-235b-a22b").function({inputTranscriptPath: summarizePath, documentTypePath: "a", language: "en"}).then(resp => {
-                done("Didnt crash")
-            }).catch(err => {
-                done()
-            })
-        })
-    })
-
-    describe("Audio Snippet", function() {
-        this.slow(1000)
-        this.timeout(5000)
-
-        // transcriptPath = "A:\\programing\\@projects\\video2document\\storage\\transcriptionSummaries\\testvideo-1765900665001.json"
-        // audiopath = "A:\\programing\\@projects\\video2document\\storage\\audio\\testvideo.mp3"
-
-
-        it("Audio Snipper Generator", function (done){
-            mapFunctions.get("extract-speaker-snippets").function({audioPath: audiopath, jsonPath: summarizePath }).then(resp => {
-                speakers = resp
-                done()
-            }).catch(err => {
-                done(err)
-            })
-        })
-
-        it("Audio Snipper Generator (Nonexistant Transcript File)", function (done){
-            mapFunctions.get("extract-speaker-snippets").function({audioPath: audiopath, jsonPath: "a" }).then(resp => {
-                speakers = resp
-                done("Didnt crash")
-            }).catch(err => {
-                done()
-            })
-        })
-
-        it("Audio Snipper Generator (Nonexistant Audio File)", function (done){
-            mapFunctions.get("extract-speaker-snippets").function({audioPath: "a", jsonPath: summarizePath }).then(resp => {
-                speakers = resp
-                done("Didnt crash")
-            }).catch(err => {
-                done()
+                throw err
             })
         })
     })