mirror of
https://github.com/OpenSquawk/OpenSquawk
synced 2026-08-01 06:06:05 +08:00
kann ein bisschen normalized sprechen
This commit is contained in:
itsrubberduck
161
server/api/atc/generate.post.ts
Normal file
161
server/api/atc/generate.post.ts
Normal file
@@ -0,0 +1,161 @@
|
||||
import { createError, getQuery } from "h3";
|
||||
import {openai, LLM_MODEL, TTS_MODEL, atcSeedPrompt, normalizeATC} from "../../utils/openai";
|
||||
import { writeFile, readFile, rm } from "node:fs/promises";
|
||||
import { randomUUID } from "node:crypto";
|
||||
import { tmpdir } from "node:os";
|
||||
import { join } from "node:path";
|
||||
import { execFile } from "node:child_process";
|
||||
|
||||
// ffmpeg-Filter je Qualitätsstufe (1..5) -> final label [out]
|
||||
function buildRadioFilter(level: number) {
|
||||
const L = Math.max(1, Math.min(5, Math.floor(level || 4)));
|
||||
const dropout = (period: number, dur: number) =>
|
||||
`volume=enable='lt(mod(t\\,${period})\\,${dur})':volume=0`;
|
||||
const band = (hp: number, lp: number, gain = 6) =>
|
||||
`highpass=f=${hp},lowpass=f=${lp},compand=attacks=0.02:decays=0.25:points=-80/-900|-70/-20|0/-10|20/-8:gain=${gain}`;
|
||||
const tail = `aecho=0.6:0.7:8:0.08,acompressor=threshold=0.6:ratio=6:attack=20:release=200`;
|
||||
|
||||
// Hilfsbausteine als vollständige Kettenglieder mit Semikolons/Labels
|
||||
const mkCrushMix = (bits: number, mix = 0.25, inLabel = "pre", outLabel = "mix1") => [
|
||||
`[${inLabel}]asplit=2[clean][toCrush]`,
|
||||
`[toCrush]acrusher=bits=${bits}:mix=1[crushed]`,
|
||||
`[clean][crushed]amix=inputs=2:weights=1 ${mix}:duration=shortest[${outLabel}]`,
|
||||
];
|
||||
|
||||
const mkNoiseMix = (amp: number, inLabel = "mix1", outLabel = "mix2") => [
|
||||
`anoisesrc=color=white:amplitude=${amp}[ns]`,
|
||||
`[${inLabel}][ns]amix=inputs=2:weights=1 0.25:duration=shortest[${outLabel}]`,
|
||||
];
|
||||
|
||||
const mkDropTail = (inLabel: string, period: number | null, dur: number | null) => {
|
||||
const chain = [];
|
||||
if (period && dur) chain.push(`[${inLabel}]${dropout(period, dur)}[drop]`);
|
||||
const src = period && dur ? "drop" : inLabel;
|
||||
chain.push(`[${src}]${tail}[out]`);
|
||||
return chain;
|
||||
};
|
||||
|
||||
if (L === 5) {
|
||||
// Dein Original, nur gelabelt bis [out]
|
||||
return [
|
||||
`[0:a]${band(300,3400)},volume=1.2[a]`,
|
||||
`anoisesrc=color=white:amplitude=0.02[ns]`,
|
||||
`[a][ns]amix=inputs=2:weights=1 0.25:duration=shortest[mix]`,
|
||||
`[mix]${tail}[out]`,
|
||||
].join(";");
|
||||
}
|
||||
|
||||
if (L === 4) {
|
||||
return [
|
||||
`[0:a]${band(320,3300)},volume=1.15[pre]`,
|
||||
...mkNoiseMix(0.03, "pre", "mix"),
|
||||
`[mix]${tail}[out]`,
|
||||
].join(";");
|
||||
}
|
||||
|
||||
if (L === 3) {
|
||||
return [
|
||||
`[0:a]${band(350,3200)},volume=1.1[pre]`,
|
||||
...mkCrushMix(12, 0.22, "pre", "mix1"),
|
||||
...mkNoiseMix(0.05, "mix1", "mix2"),
|
||||
...mkDropTail("mix2", 6, 0.06),
|
||||
].join(";");
|
||||
}
|
||||
|
||||
if (L === 2) {
|
||||
return [
|
||||
`[0:a]${band(400,3000)},volume=1.05[pre]`,
|
||||
...mkCrushMix(10, 0.32, "pre", "mix1"),
|
||||
// mehr Noise-Gewichtung beim zweiten Mix
|
||||
`anoisesrc=color=white:amplitude=0.08[ns]`,
|
||||
`[mix1][ns]amix=inputs=2:weights=1 0.6:duration=shortest[mix2]`,
|
||||
...mkDropTail("mix2", 4.5, 0.12),
|
||||
].join(";");
|
||||
}
|
||||
|
||||
// L === 1
|
||||
return [
|
||||
`[0:a]${band(500,2600,5)},volume=1.0[pre]`,
|
||||
...mkCrushMix(8, 0.45, "pre", "mix1"),
|
||||
`anoisesrc=color=white:amplitude=0.12[ns]`,
|
||||
`[mix1][ns]amix=inputs=2:weights=1 0.8:duration=shortest[mix2]`,
|
||||
...mkDropTail("mix2", 3.5, 0.2),
|
||||
].join(";");
|
||||
}
|
||||
|
||||
// Funk-Effekt mit ffmpeg (fix: -map [out], kein "[post]?0:a")
|
||||
async function applyRadioEffect(input: string, output: string, level = 4) {
|
||||
const filter = buildRadioFilter(level);
|
||||
await new Promise<void>((res, rej) =>
|
||||
execFile(
|
||||
"ffmpeg",
|
||||
["-y", "-i", input, "-filter_complex", filter, "-map", "[out]", "-ar", "16000", output],
|
||||
(err, _o, stderr) => (err ? rej(new Error(stderr || String(err))) : res())
|
||||
)
|
||||
);
|
||||
}
|
||||
|
||||
export default defineEventHandler(async (event) => {
|
||||
// level aus query (?level=1..5), default 4
|
||||
const { level } = getQuery(event);
|
||||
const lvl = Math.max(1, Math.min(5, parseInt(String(level ?? "4"), 10) || 4));
|
||||
|
||||
// 1) ATC-Text erzeugen (ohne Pilot-Input)
|
||||
const scenario = {
|
||||
airport: "EDDF",
|
||||
aircraft: "A320",
|
||||
type: "IFR",
|
||||
stand: "V155",
|
||||
dep: "EHAM",
|
||||
sid: "MARUN 7F",
|
||||
squawk: "4723",
|
||||
freq: "121.800",
|
||||
runway: "25R",
|
||||
phase: "taxi",
|
||||
notes: "Taxiway N closed between N2–N4",
|
||||
};
|
||||
|
||||
// const resp = await openai.responses.create({
|
||||
// model: LLM_MODEL,
|
||||
// input: atcSeedPrompt(scenario),
|
||||
// });
|
||||
// const atcText = (resp.output_text || "").trim();
|
||||
const atcText = "DLH39A taxi to RWY 25R via V A, hold short of RWY 25R.";
|
||||
if (!atcText) throw createError({ statusCode: 500, statusMessage: "LLM empty" });
|
||||
|
||||
const normalized = normalizeATC(atcText);
|
||||
if (!normalized) throw createError({ statusCode: 500, statusMessage: "ATC text empty" });
|
||||
console.log("ATC Text (normalized):", normalized);
|
||||
|
||||
// return { atcText: normalized, level: lvl };
|
||||
|
||||
// 2) TTS (clean)
|
||||
const cleanPath = join(tmpdir(), `tts-${randomUUID()}.wav`);
|
||||
const radioPath = join(tmpdir(), `radio-${randomUUID()}.wav`);
|
||||
|
||||
const tts = await openai.audio.speech.create({
|
||||
model: TTS_MODEL,
|
||||
voice: "alloy",
|
||||
format: "wav",
|
||||
input: normalized,
|
||||
});
|
||||
await writeFile(cleanPath, Buffer.from(await tts.arrayBuffer()));
|
||||
|
||||
// 3) Funk-Effekt (mit stufe)
|
||||
await applyRadioEffect(cleanPath, radioPath, lvl);
|
||||
|
||||
// 4) Payload zurück (Debug: Text + beide Audios)
|
||||
const cleanB64 = (await readFile(cleanPath)).toString("base64");
|
||||
const radioB64 = (await readFile(radioPath)).toString("base64");
|
||||
rm(cleanPath).catch(() => {});
|
||||
rm(radioPath).catch(() => {});
|
||||
|
||||
return {
|
||||
atcText,
|
||||
level: lvl,
|
||||
audio: {
|
||||
clean: { mime: "audio/wav", base64: cleanB64 },
|
||||
radio: { mime: "audio/wav", base64: radioB64 },
|
||||
},
|
||||
};
|
||||
});
|
||||
@@ -1,62 +1,56 @@
|
||||
import { readMultipartFormData, createError } from "h3";
|
||||
import { LLM_MODEL, TTS_MODEL, atcReplyPrompt, openai } from "../../utils/openai";
|
||||
import { openai, LLM_MODEL, TTS_MODEL, atcReplyPrompt } from "../../utils/openai";
|
||||
import { applyRadioEffect } from "../../utils/radio";
|
||||
import { writeFile, rm, readFile } from "node:fs/promises";
|
||||
import { randomUUID } from "node:crypto";
|
||||
import { tmpdir } from "node:os";
|
||||
import { join } from "node:path";
|
||||
import { toFile } from "openai/uploads"; // ⟵ NEU
|
||||
import { toFile } from "openai/uploads";
|
||||
|
||||
export default defineEventHandler(async (event) => {
|
||||
const parts = await readMultipartFormData(event);
|
||||
if (!parts) throw createError({ statusCode: 400, statusMessage: "No form-data" });
|
||||
|
||||
const mode = (parts.find(p => p.name === "mode")?.data?.toString("utf8") || "text").toLowerCase();
|
||||
// Default ohne Text-Input: immer TTS
|
||||
const audio = parts.find(p => p.type && p.data);
|
||||
if (!audio) throw createError({ statusCode: 400, statusMessage: "No audio file" });
|
||||
|
||||
// ✅ saubere Datei-Erzeugung
|
||||
const file = await toFile(
|
||||
audio.data,
|
||||
audio.filename || "ptt.webm",
|
||||
{ type: audio.type || "audio/webm" }
|
||||
);
|
||||
|
||||
// 1) Transkribieren
|
||||
// 1) Transkription (Whisper)
|
||||
const tr = await openai.audio.transcriptions.create({ model: "whisper-1", file });
|
||||
const pilotText = tr.text?.trim() || "";
|
||||
const pilotText = (tr.text || "").trim();
|
||||
if (!pilotText) throw createError({ statusCode: 400, statusMessage: "Empty transcription" });
|
||||
|
||||
// 2) ATC-Antwort
|
||||
// 2) ATC-Antwort (LLM)
|
||||
const resp = await openai.responses.create({
|
||||
model: LLM_MODEL,
|
||||
input: atcReplyPrompt(pilotText),
|
||||
});
|
||||
const replyText = resp.output_text?.trim() || "";
|
||||
const replyText = (resp.output_text || "").trim();
|
||||
if (!replyText) throw createError({ statusCode: 500, statusMessage: "LLM empty" });
|
||||
|
||||
// 3) Optional TTS + Funk
|
||||
if (mode === "tts") {
|
||||
const clean = join(tmpdir(), `tts-${randomUUID()}.wav`);
|
||||
const radio = join(tmpdir(), `radio-${randomUUID()}.wav`);
|
||||
// 3) TTS + Funk-Effekt
|
||||
const clean = join(tmpdir(), `tts-${randomUUID()}.wav`);
|
||||
const radio = join(tmpdir(), `radio-${randomUUID()}.wav`);
|
||||
|
||||
const tts = await openai.audio.speech.create({
|
||||
model: TTS_MODEL,
|
||||
voice: "alloy",
|
||||
format: "wav",
|
||||
input: replyText,
|
||||
});
|
||||
const tts = await openai.audio.speech.create({
|
||||
model: TTS_MODEL,
|
||||
voice: "alloy",
|
||||
format: "wav",
|
||||
input: replyText,
|
||||
});
|
||||
await writeFile(clean, Buffer.from(await tts.arrayBuffer()));
|
||||
await applyRadioEffect(clean, radio);
|
||||
|
||||
await writeFile(clean, Buffer.from(await tts.arrayBuffer()));
|
||||
await applyRadioEffect(clean, radio);
|
||||
const data = await readFile(radio);
|
||||
const b64 = Buffer.from(data).toString("base64");
|
||||
rm(clean).catch(() => {});
|
||||
rm(radio).catch(() => {});
|
||||
|
||||
const data = await readFile(radio);
|
||||
const b64 = Buffer.from(data).toString("base64");
|
||||
rm(clean).catch(() => {});
|
||||
rm(radio).catch(() => {});
|
||||
|
||||
return { pilotText, replyText, audio: { mime: "audio/wav", base64: b64 } };
|
||||
}
|
||||
|
||||
return { pilotText, replyText };
|
||||
return { pilotText, replyText, audio: { mime: "audio/wav", base64: b64 } };
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user