plotagon-director 0.6.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- plotagon_director-0.6.0/LICENSE +21 -0
- plotagon_director-0.6.0/PKG-INFO +143 -0
- plotagon_director-0.6.0/README.md +117 -0
- plotagon_director-0.6.0/plotagon_director/__init__.py +3 -0
- plotagon_director-0.6.0/plotagon_director/audio/__init__.py +0 -0
- plotagon_director-0.6.0/plotagon_director/audio/phonemes.py +70 -0
- plotagon_director-0.6.0/plotagon_director/audio/tts.py +163 -0
- plotagon_director-0.6.0/plotagon_director/cli.py +308 -0
- plotagon_director-0.6.0/plotagon_director/core/__init__.py +0 -0
- plotagon_director-0.6.0/plotagon_director/core/archive.py +35 -0
- plotagon_director-0.6.0/plotagon_director/core/decompiler.py +120 -0
- plotagon_director-0.6.0/plotagon_director/core/diffplot.py +89 -0
- plotagon_director-0.6.0/plotagon_director/core/doctor.py +67 -0
- plotagon_director-0.6.0/plotagon_director/core/plotdoc.py +178 -0
- plotagon_director-0.6.0/plotagon_director/core/validator.py +165 -0
- plotagon_director-0.6.0/plotagon_director/library/__init__.py +14 -0
- plotagon_director-0.6.0/plotagon_director/library/builder.py +153 -0
- plotagon_director-0.6.0/plotagon_director/library/data/library.json +4819 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/PKG-INFO +143 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/SOURCES.txt +28 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/dependency_links.txt +1 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/entry_points.txt +2 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/requires.txt +12 -0
- plotagon_director-0.6.0/plotagon_director.egg-info/top_level.txt +1 -0
- plotagon_director-0.6.0/pyproject.toml +41 -0
- plotagon_director-0.6.0/setup.cfg +4 -0
- plotagon_director-0.6.0/tests/test_actions.py +47 -0
- plotagon_director-0.6.0/tests/test_cli_extras.py +81 -0
- plotagon_director-0.6.0/tests/test_music_camera.py +66 -0
- plotagon_director-0.6.0/tests/test_validator.py +87 -0
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Mohamed Amine Ben Mallessa
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
|
@@ -0,0 +1,143 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: plotagon-director
|
|
3
|
+
Version: 0.6.0
|
|
4
|
+
Summary: Génération de films Plotagon Studio Desktop (.plot) avec voix Edge TTS et synchro labiale
|
|
5
|
+
Author: Mohamed Amine Ben Mallessa
|
|
6
|
+
License: MIT
|
|
7
|
+
Project-URL: LinkedIn, https://www.linkedin.com/in/benmallessa/
|
|
8
|
+
Keywords: plotagon,tts,video,animation
|
|
9
|
+
Classifier: Programming Language :: Python :: 3
|
|
10
|
+
Classifier: License :: OSI Approved :: MIT License
|
|
11
|
+
Classifier: Operating System :: OS Independent
|
|
12
|
+
Requires-Python: >=3.10
|
|
13
|
+
Description-Content-Type: text/markdown
|
|
14
|
+
License-File: LICENSE
|
|
15
|
+
Requires-Dist: pyyaml
|
|
16
|
+
Requires-Dist: edge-tts
|
|
17
|
+
Requires-Dist: soundfile
|
|
18
|
+
Requires-Dist: numpy
|
|
19
|
+
Requires-Dist: truststore
|
|
20
|
+
Provides-Extra: library
|
|
21
|
+
Requires-Dist: UnityPy; extra == "library"
|
|
22
|
+
Provides-Extra: dev
|
|
23
|
+
Requires-Dist: pytest; extra == "dev"
|
|
24
|
+
Requires-Dist: UnityPy; extra == "dev"
|
|
25
|
+
Dynamic: license-file
|
|
26
|
+
|
|
27
|
+
# Plotagon Director
|
|
28
|
+
|
|
29
|
+
Génération de films [Plotagon Studio Desktop](https://www.plotagon.com) de
|
|
30
|
+
bout en bout : écriture du scénario en YAML, validation contre la bibliothèque
|
|
31
|
+
de ressources réelle, génération d'un fichier `.plot` importable dans
|
|
32
|
+
l'application officielle — avec voix custom Edge TTS et synchro labiale.
|
|
33
|
+
|
|
34
|
+
Le format `.plot` a été entièrement rétro-ingénieré et confirmé par analyse
|
|
35
|
+
d'exports officiels (v1.11.0) : ZIP contenant un document JSON `.plotdoc`
|
|
36
|
+
(liste plate d'instructions `scene`, `dialogue`, `music`, `action`) et, par
|
|
37
|
+
réplique vocalisée, un WAV mono 44,1 kHz + un fichier `.phonemes` (protobuf)
|
|
38
|
+
pour la synchro labiale. Spécification complète : `docs/PLOT_FORMAT.md`.
|
|
39
|
+
|
|
40
|
+
## Installation
|
|
41
|
+
|
|
42
|
+
```bash
|
|
43
|
+
pip install -e . # depuis un clone: commande `plotagon-director`
|
|
44
|
+
# ou directement depuis GitHub :
|
|
45
|
+
pip install git+https://github.com/mohamed-amine-ben-mallessa/plotagon-director.git
|
|
46
|
+
```
|
|
47
|
+
|
|
48
|
+
La bibliothèque de ressources (`library.json`) est embarquée dans le package :
|
|
49
|
+
le générateur fonctionne sans installation Plotagon. L'extra `[library]`
|
|
50
|
+
(`pip install -e .[library]`, ajoute UnityPy) n'est nécessaire que pour
|
|
51
|
+
**reconstruire** la bibliothèque depuis une installation Plotagon.
|
|
52
|
+
|
|
53
|
+
## Usage
|
|
54
|
+
|
|
55
|
+
```bash
|
|
56
|
+
# Écrire un scénario
|
|
57
|
+
plotagon-director new "Mon film" # squelette commenté
|
|
58
|
+
plotagon-director list scenes --filter cafe # aussi: characters, expressions,
|
|
59
|
+
plotagon-director list actions # music, sounds, voices
|
|
60
|
+
plotagon-director positions restaurants.cafe # slots d'acteurs d'une scène
|
|
61
|
+
plotagon-director voices fr # voix Edge TTS disponibles
|
|
62
|
+
plotagon-director say "Bonjour" fr-FR-DeniseNeural # écouter une voix
|
|
63
|
+
|
|
64
|
+
# Valider puis générer
|
|
65
|
+
plotagon-director validate mon-film.yaml [--strict]
|
|
66
|
+
plotagon-director build mon-film.yaml mon-film.plot
|
|
67
|
+
# --no-voices : hors ligne | --srt subs.srt : sous-titres
|
|
68
|
+
# --no-cache : resynthétise tout (sinon cache TTS automatique)
|
|
69
|
+
plotagon-director batch scenarios/ sorties/ # tous les .yaml d'un dossier
|
|
70
|
+
|
|
71
|
+
# Importer mon-film.plot dans Plotagon Studio (menu import)
|
|
72
|
+
|
|
73
|
+
# Analyse / rétro-ingénierie
|
|
74
|
+
plotagon-director parse export.plot # JSON brut d'un .plot
|
|
75
|
+
plotagon-director decompile export.plot film.yaml # .plot -> YAML éditable
|
|
76
|
+
plotagon-director diff avant.plot apres.plot # isoler un format inconnu
|
|
77
|
+
plotagon-director doctor # diagnostic environnement
|
|
78
|
+
|
|
79
|
+
# (Reconstruire la bibliothèque depuis une installation Plotagon)
|
|
80
|
+
python -m plotagon_director.library.builder [chemin/builtincontentmanifest.json]
|
|
81
|
+
```
|
|
82
|
+
|
|
83
|
+
## Exemple de scénario
|
|
84
|
+
|
|
85
|
+
```yaml
|
|
86
|
+
project:
|
|
87
|
+
title: "Conversation au café"
|
|
88
|
+
|
|
89
|
+
cast:
|
|
90
|
+
- ref: paul
|
|
91
|
+
character: news.paul
|
|
92
|
+
name: "Paul"
|
|
93
|
+
voice: fr-FR-HenriNeural # optionnel : voix Edge TTS (+ rate/pitch)
|
|
94
|
+
- ref: lucy
|
|
95
|
+
character: news.lucy
|
|
96
|
+
name: "Lucy"
|
|
97
|
+
voice: fr-FR-DeniseNeural
|
|
98
|
+
|
|
99
|
+
scenes:
|
|
100
|
+
- scene: restaurants.cafe
|
|
101
|
+
music: music.corny # optionnel (music.stopmusic pour arrêter)
|
|
102
|
+
actors:
|
|
103
|
+
- ref: paul
|
|
104
|
+
position: Table1 # `plotagon-director positions restaurants.cafe`
|
|
105
|
+
- ref: lucy
|
|
106
|
+
position: Table2
|
|
107
|
+
dialogue:
|
|
108
|
+
- actor: paul
|
|
109
|
+
expression: happy
|
|
110
|
+
text: "Bonjour Lucy !"
|
|
111
|
+
camera: wide shot # optionnel : 10 types (docs/PLOT_FORMAT.md)
|
|
112
|
+
- actor: paul # interaction à 2 personnages
|
|
113
|
+
action: handshake
|
|
114
|
+
target: lucy
|
|
115
|
+
```
|
|
116
|
+
|
|
117
|
+
Bibliothèque embarquée : 205 scènes (953 positions), 53 personnages,
|
|
118
|
+
88 expressions, 15 actions, 38 musiques, 132 sons.
|
|
119
|
+
|
|
120
|
+
## Architecture
|
|
121
|
+
|
|
122
|
+
- `plotagon_director/core/` — modèle `.plotdoc`, conteneur ZIP, validateur,
|
|
123
|
+
décompilateur, diff
|
|
124
|
+
- `plotagon_director/audio/` — Edge TTS → WAV conforme (cache + parallèle),
|
|
125
|
+
phonèmes/lip-sync
|
|
126
|
+
- `plotagon_director/library/` — builder du catalogue + `data/library.json`
|
|
127
|
+
- `docs/` — rétro-ingénierie du format (`PLOT_FORMAT.md`), API interne de
|
|
128
|
+
l'appli (`APP_JS_ANALYSIS.md`), état des connaissances (`FINDINGS.md`)
|
|
129
|
+
- `SKILL.md` — utilisation comme skill Claude Code
|
|
130
|
+
|
|
131
|
+
## Principes
|
|
132
|
+
|
|
133
|
+
- Aucun mod, aucun binaire tiers, aucun credential.
|
|
134
|
+
- Installation officielle en lecture seule.
|
|
135
|
+
- Seuls les éléments confirmés par un export officiel sont générés ; le
|
|
136
|
+
reste (effets visuels, sons importés) est documenté UNKNOWN dans
|
|
137
|
+
`docs/FINDINGS.md`.
|
|
138
|
+
|
|
139
|
+
## Tests
|
|
140
|
+
|
|
141
|
+
```bash
|
|
142
|
+
python -m pytest tests/ -q
|
|
143
|
+
```
|
|
@@ -0,0 +1,117 @@
|
|
|
1
|
+
# Plotagon Director
|
|
2
|
+
|
|
3
|
+
Génération de films [Plotagon Studio Desktop](https://www.plotagon.com) de
|
|
4
|
+
bout en bout : écriture du scénario en YAML, validation contre la bibliothèque
|
|
5
|
+
de ressources réelle, génération d'un fichier `.plot` importable dans
|
|
6
|
+
l'application officielle — avec voix custom Edge TTS et synchro labiale.
|
|
7
|
+
|
|
8
|
+
Le format `.plot` a été entièrement rétro-ingénieré et confirmé par analyse
|
|
9
|
+
d'exports officiels (v1.11.0) : ZIP contenant un document JSON `.plotdoc`
|
|
10
|
+
(liste plate d'instructions `scene`, `dialogue`, `music`, `action`) et, par
|
|
11
|
+
réplique vocalisée, un WAV mono 44,1 kHz + un fichier `.phonemes` (protobuf)
|
|
12
|
+
pour la synchro labiale. Spécification complète : `docs/PLOT_FORMAT.md`.
|
|
13
|
+
|
|
14
|
+
## Installation
|
|
15
|
+
|
|
16
|
+
```bash
|
|
17
|
+
pip install -e . # depuis un clone: commande `plotagon-director`
|
|
18
|
+
# ou directement depuis GitHub :
|
|
19
|
+
pip install git+https://github.com/mohamed-amine-ben-mallessa/plotagon-director.git
|
|
20
|
+
```
|
|
21
|
+
|
|
22
|
+
La bibliothèque de ressources (`library.json`) est embarquée dans le package :
|
|
23
|
+
le générateur fonctionne sans installation Plotagon. L'extra `[library]`
|
|
24
|
+
(`pip install -e .[library]`, ajoute UnityPy) n'est nécessaire que pour
|
|
25
|
+
**reconstruire** la bibliothèque depuis une installation Plotagon.
|
|
26
|
+
|
|
27
|
+
## Usage
|
|
28
|
+
|
|
29
|
+
```bash
|
|
30
|
+
# Écrire un scénario
|
|
31
|
+
plotagon-director new "Mon film" # squelette commenté
|
|
32
|
+
plotagon-director list scenes --filter cafe # aussi: characters, expressions,
|
|
33
|
+
plotagon-director list actions # music, sounds, voices
|
|
34
|
+
plotagon-director positions restaurants.cafe # slots d'acteurs d'une scène
|
|
35
|
+
plotagon-director voices fr # voix Edge TTS disponibles
|
|
36
|
+
plotagon-director say "Bonjour" fr-FR-DeniseNeural # écouter une voix
|
|
37
|
+
|
|
38
|
+
# Valider puis générer
|
|
39
|
+
plotagon-director validate mon-film.yaml [--strict]
|
|
40
|
+
plotagon-director build mon-film.yaml mon-film.plot
|
|
41
|
+
# --no-voices : hors ligne | --srt subs.srt : sous-titres
|
|
42
|
+
# --no-cache : resynthétise tout (sinon cache TTS automatique)
|
|
43
|
+
plotagon-director batch scenarios/ sorties/ # tous les .yaml d'un dossier
|
|
44
|
+
|
|
45
|
+
# Importer mon-film.plot dans Plotagon Studio (menu import)
|
|
46
|
+
|
|
47
|
+
# Analyse / rétro-ingénierie
|
|
48
|
+
plotagon-director parse export.plot # JSON brut d'un .plot
|
|
49
|
+
plotagon-director decompile export.plot film.yaml # .plot -> YAML éditable
|
|
50
|
+
plotagon-director diff avant.plot apres.plot # isoler un format inconnu
|
|
51
|
+
plotagon-director doctor # diagnostic environnement
|
|
52
|
+
|
|
53
|
+
# (Reconstruire la bibliothèque depuis une installation Plotagon)
|
|
54
|
+
python -m plotagon_director.library.builder [chemin/builtincontentmanifest.json]
|
|
55
|
+
```
|
|
56
|
+
|
|
57
|
+
## Exemple de scénario
|
|
58
|
+
|
|
59
|
+
```yaml
|
|
60
|
+
project:
|
|
61
|
+
title: "Conversation au café"
|
|
62
|
+
|
|
63
|
+
cast:
|
|
64
|
+
- ref: paul
|
|
65
|
+
character: news.paul
|
|
66
|
+
name: "Paul"
|
|
67
|
+
voice: fr-FR-HenriNeural # optionnel : voix Edge TTS (+ rate/pitch)
|
|
68
|
+
- ref: lucy
|
|
69
|
+
character: news.lucy
|
|
70
|
+
name: "Lucy"
|
|
71
|
+
voice: fr-FR-DeniseNeural
|
|
72
|
+
|
|
73
|
+
scenes:
|
|
74
|
+
- scene: restaurants.cafe
|
|
75
|
+
music: music.corny # optionnel (music.stopmusic pour arrêter)
|
|
76
|
+
actors:
|
|
77
|
+
- ref: paul
|
|
78
|
+
position: Table1 # `plotagon-director positions restaurants.cafe`
|
|
79
|
+
- ref: lucy
|
|
80
|
+
position: Table2
|
|
81
|
+
dialogue:
|
|
82
|
+
- actor: paul
|
|
83
|
+
expression: happy
|
|
84
|
+
text: "Bonjour Lucy !"
|
|
85
|
+
camera: wide shot # optionnel : 10 types (docs/PLOT_FORMAT.md)
|
|
86
|
+
- actor: paul # interaction à 2 personnages
|
|
87
|
+
action: handshake
|
|
88
|
+
target: lucy
|
|
89
|
+
```
|
|
90
|
+
|
|
91
|
+
Bibliothèque embarquée : 205 scènes (953 positions), 53 personnages,
|
|
92
|
+
88 expressions, 15 actions, 38 musiques, 132 sons.
|
|
93
|
+
|
|
94
|
+
## Architecture
|
|
95
|
+
|
|
96
|
+
- `plotagon_director/core/` — modèle `.plotdoc`, conteneur ZIP, validateur,
|
|
97
|
+
décompilateur, diff
|
|
98
|
+
- `plotagon_director/audio/` — Edge TTS → WAV conforme (cache + parallèle),
|
|
99
|
+
phonèmes/lip-sync
|
|
100
|
+
- `plotagon_director/library/` — builder du catalogue + `data/library.json`
|
|
101
|
+
- `docs/` — rétro-ingénierie du format (`PLOT_FORMAT.md`), API interne de
|
|
102
|
+
l'appli (`APP_JS_ANALYSIS.md`), état des connaissances (`FINDINGS.md`)
|
|
103
|
+
- `SKILL.md` — utilisation comme skill Claude Code
|
|
104
|
+
|
|
105
|
+
## Principes
|
|
106
|
+
|
|
107
|
+
- Aucun mod, aucun binaire tiers, aucun credential.
|
|
108
|
+
- Installation officielle en lecture seule.
|
|
109
|
+
- Seuls les éléments confirmés par un export officiel sont générés ; le
|
|
110
|
+
reste (effets visuels, sons importés) est documenté UNKNOWN dans
|
|
111
|
+
`docs/FINDINGS.md`.
|
|
112
|
+
|
|
113
|
+
## Tests
|
|
114
|
+
|
|
115
|
+
```bash
|
|
116
|
+
python -m pytest tests/ -q
|
|
117
|
+
```
|
|
File without changes
|
|
@@ -0,0 +1,70 @@
|
|
|
1
|
+
"""Fichiers .phonemes (synchro labiale) au format de Plotagon Studio.
|
|
2
|
+
|
|
3
|
+
Format CONFIRME par decodage d'un export officiel : messages protobuf
|
|
4
|
+
concatenes, un par trame de 2048 echantillons (~46.4 ms a 44100 Hz) :
|
|
5
|
+
|
|
6
|
+
champ 1 (tag 0x0d, float32 LE) : debut en secondes (omis si 0)
|
|
7
|
+
champ 2 (tag 0x15, float32 LE) : fin en secondes
|
|
8
|
+
champ 3 (tag 0x1a, string) : phoneme
|
|
9
|
+
|
|
10
|
+
Phonemes observes : sil, ae, d, eh, l, m, n, ow, uu, uw.
|
|
11
|
+
"""
|
|
12
|
+
|
|
13
|
+
import io
|
|
14
|
+
import struct
|
|
15
|
+
import wave
|
|
16
|
+
|
|
17
|
+
FRAME_SAMPLES = 2048
|
|
18
|
+
|
|
19
|
+
# Cycle de voyelles pour la synchro labiale approximative (amplitude-based).
|
|
20
|
+
_MOUTH_CYCLE = ["eh", "ae", "ow", "eh", "uw", "ae"]
|
|
21
|
+
|
|
22
|
+
|
|
23
|
+
def phonemes_from_wav(wav_bytes):
|
|
24
|
+
"""Synchro labiale approximative par enveloppe d'amplitude.
|
|
25
|
+
|
|
26
|
+
L'appli genere les phonemes par analyse du signal ; ici on approxime :
|
|
27
|
+
trames silencieuses -> "sil", trames parlees -> cycle de voyelles.
|
|
28
|
+
"""
|
|
29
|
+
import numpy as np
|
|
30
|
+
|
|
31
|
+
with wave.open(io.BytesIO(wav_bytes)) as w:
|
|
32
|
+
rate = w.getframerate()
|
|
33
|
+
pcm = np.frombuffer(w.readframes(w.getnframes()), dtype="<i2")
|
|
34
|
+
|
|
35
|
+
records = []
|
|
36
|
+
n_frames = max(1, (len(pcm) + FRAME_SAMPLES - 1) // FRAME_SAMPLES)
|
|
37
|
+
rms = np.array([
|
|
38
|
+
np.sqrt(np.mean(
|
|
39
|
+
pcm[i * FRAME_SAMPLES:(i + 1) * FRAME_SAMPLES].astype("f8") ** 2
|
|
40
|
+
)) if len(pcm[i * FRAME_SAMPLES:(i + 1) * FRAME_SAMPLES]) else 0.0
|
|
41
|
+
for i in range(n_frames)
|
|
42
|
+
])
|
|
43
|
+
threshold = max(300.0, float(rms.max()) * 0.08)
|
|
44
|
+
|
|
45
|
+
mouth = 0
|
|
46
|
+
for i in range(n_frames):
|
|
47
|
+
start = i * FRAME_SAMPLES / rate
|
|
48
|
+
end = min((i + 1) * FRAME_SAMPLES, len(pcm)) / rate
|
|
49
|
+
if rms[i] >= threshold:
|
|
50
|
+
label = _MOUTH_CYCLE[(mouth // 2) % len(_MOUTH_CYCLE)]
|
|
51
|
+
mouth += 1
|
|
52
|
+
else:
|
|
53
|
+
label = "sil"
|
|
54
|
+
mouth = 0
|
|
55
|
+
records.append((start, end, label))
|
|
56
|
+
return records
|
|
57
|
+
|
|
58
|
+
|
|
59
|
+
def encode_phonemes(records):
|
|
60
|
+
"""Encode les enregistrements au format protobuf observe."""
|
|
61
|
+
out = bytearray()
|
|
62
|
+
for start, end, label in records:
|
|
63
|
+
body = bytearray()
|
|
64
|
+
if start > 0.0:
|
|
65
|
+
body += b"\x0d" + struct.pack("<f", start)
|
|
66
|
+
body += b"\x15" + struct.pack("<f", end)
|
|
67
|
+
raw = label.encode()
|
|
68
|
+
body += b"\x1a" + bytes([len(raw)]) + raw
|
|
69
|
+
out += b"\x0a" + bytes([len(body)]) + body
|
|
70
|
+
return bytes(out)
|
|
@@ -0,0 +1,163 @@
|
|
|
1
|
+
"""Synthese vocale Edge TTS au format audio de Plotagon Studio.
|
|
2
|
+
|
|
3
|
+
Format CONFIRME (export officiel 1.11.0) : WAV PCM mono 44100 Hz 16 bits.
|
|
4
|
+
|
|
5
|
+
Deux optimisations pour l'iteration rapide :
|
|
6
|
+
- cache disque (%LOCALAPPDATA%/plotagon-director/tts-cache) par hash de
|
|
7
|
+
(texte, voix, rate, pitch) — un rebuild sans changement ne touche pas
|
|
8
|
+
au reseau ;
|
|
9
|
+
- synthese parallele (asyncio) pour les films a nombreuses repliques.
|
|
10
|
+
"""
|
|
11
|
+
|
|
12
|
+
import asyncio
|
|
13
|
+
import hashlib
|
|
14
|
+
import io
|
|
15
|
+
import os
|
|
16
|
+
import wave
|
|
17
|
+
from pathlib import Path
|
|
18
|
+
|
|
19
|
+
TARGET_RATE = 44100
|
|
20
|
+
CONCURRENCY = 4
|
|
21
|
+
|
|
22
|
+
CACHE_DIR = (
|
|
23
|
+
Path(os.environ.get("LOCALAPPDATA", str(Path.home())))
|
|
24
|
+
/ "plotagon-director" / "tts-cache"
|
|
25
|
+
)
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
def _inject_truststore():
|
|
29
|
+
"""Fait utiliser a Python le magasin de certificats Windows.
|
|
30
|
+
|
|
31
|
+
Indispensable derriere un antivirus qui intercepte le TLS (ex. Norton).
|
|
32
|
+
"""
|
|
33
|
+
try:
|
|
34
|
+
import truststore
|
|
35
|
+
truststore.inject_into_ssl()
|
|
36
|
+
except ImportError:
|
|
37
|
+
pass
|
|
38
|
+
|
|
39
|
+
|
|
40
|
+
def _cache_path(text, voice, rate, pitch):
|
|
41
|
+
key = "%s|%s|%s|%s" % (text, voice, rate or "", pitch or "")
|
|
42
|
+
return CACHE_DIR / (hashlib.sha256(key.encode("utf-8")).hexdigest() + ".wav")
|
|
43
|
+
|
|
44
|
+
|
|
45
|
+
async def _fetch_mp3(text, voice, rate=None, pitch=None):
|
|
46
|
+
import edge_tts
|
|
47
|
+
|
|
48
|
+
kwargs = {}
|
|
49
|
+
if rate:
|
|
50
|
+
kwargs["rate"] = rate
|
|
51
|
+
if pitch:
|
|
52
|
+
kwargs["pitch"] = pitch
|
|
53
|
+
communicate = edge_tts.Communicate(text, voice, **kwargs)
|
|
54
|
+
buf = io.BytesIO()
|
|
55
|
+
async for chunk in communicate.stream():
|
|
56
|
+
if chunk["type"] == "audio":
|
|
57
|
+
buf.write(chunk["data"])
|
|
58
|
+
return buf.getvalue()
|
|
59
|
+
|
|
60
|
+
|
|
61
|
+
def _mp3_to_wav(mp3_bytes):
|
|
62
|
+
import numpy as np
|
|
63
|
+
import soundfile as sf
|
|
64
|
+
|
|
65
|
+
data, sr = sf.read(io.BytesIO(mp3_bytes), dtype="float32")
|
|
66
|
+
if data.ndim > 1:
|
|
67
|
+
data = data.mean(axis=1)
|
|
68
|
+
if sr != TARGET_RATE:
|
|
69
|
+
n_out = int(round(len(data) * TARGET_RATE / sr))
|
|
70
|
+
x_old = np.linspace(0.0, 1.0, num=len(data), endpoint=False)
|
|
71
|
+
x_new = np.linspace(0.0, 1.0, num=n_out, endpoint=False)
|
|
72
|
+
data = np.interp(x_new, x_old, data).astype("float32")
|
|
73
|
+
|
|
74
|
+
pcm = (np.clip(data, -1.0, 1.0) * 32767).astype("<i2")
|
|
75
|
+
out = io.BytesIO()
|
|
76
|
+
with wave.open(out, "wb") as w:
|
|
77
|
+
w.setnchannels(1)
|
|
78
|
+
w.setsampwidth(2)
|
|
79
|
+
w.setframerate(TARGET_RATE)
|
|
80
|
+
w.writeframes(pcm.tobytes())
|
|
81
|
+
return out.getvalue()
|
|
82
|
+
|
|
83
|
+
|
|
84
|
+
def synthesize_wav(text, voice, rate=None, pitch=None, use_cache=True):
|
|
85
|
+
"""Synthetise une replique et retourne des octets WAV 44.1k mono 16b."""
|
|
86
|
+
if use_cache:
|
|
87
|
+
cached = _cache_path(text, voice, rate, pitch)
|
|
88
|
+
if cached.exists():
|
|
89
|
+
return cached.read_bytes()
|
|
90
|
+
|
|
91
|
+
_inject_truststore()
|
|
92
|
+
wav = _mp3_to_wav(asyncio.run(_fetch_mp3(text, voice, rate, pitch)))
|
|
93
|
+
|
|
94
|
+
if use_cache:
|
|
95
|
+
CACHE_DIR.mkdir(parents=True, exist_ok=True)
|
|
96
|
+
cached.write_bytes(wav)
|
|
97
|
+
return wav
|
|
98
|
+
|
|
99
|
+
|
|
100
|
+
def synthesize_many(jobs, use_cache=True, progress=None):
|
|
101
|
+
"""Synthetise plusieurs repliques en parallele.
|
|
102
|
+
|
|
103
|
+
jobs : liste de dicts {"guid", "text", "voice", "rate", "pitch"}.
|
|
104
|
+
progress : callback(job, from_cache) appele apres chaque replique.
|
|
105
|
+
Retourne {guid: wav_bytes}.
|
|
106
|
+
"""
|
|
107
|
+
results = {}
|
|
108
|
+
to_fetch = []
|
|
109
|
+
for job in jobs:
|
|
110
|
+
cached = _cache_path(job["text"], job["voice"], job.get("rate"), job.get("pitch"))
|
|
111
|
+
if use_cache and cached.exists():
|
|
112
|
+
results[job["guid"]] = cached.read_bytes()
|
|
113
|
+
if progress:
|
|
114
|
+
progress(job, True)
|
|
115
|
+
else:
|
|
116
|
+
to_fetch.append(job)
|
|
117
|
+
|
|
118
|
+
if to_fetch:
|
|
119
|
+
_inject_truststore()
|
|
120
|
+
|
|
121
|
+
async def _run_all():
|
|
122
|
+
sem = asyncio.Semaphore(CONCURRENCY)
|
|
123
|
+
|
|
124
|
+
async def _one(job):
|
|
125
|
+
async with sem:
|
|
126
|
+
mp3 = await _fetch_mp3(
|
|
127
|
+
job["text"], job["voice"], job.get("rate"), job.get("pitch")
|
|
128
|
+
)
|
|
129
|
+
return job, mp3
|
|
130
|
+
|
|
131
|
+
return await asyncio.gather(*(_one(j) for j in to_fetch))
|
|
132
|
+
|
|
133
|
+
for job, mp3 in asyncio.run(_run_all()):
|
|
134
|
+
wav = _mp3_to_wav(mp3)
|
|
135
|
+
results[job["guid"]] = wav
|
|
136
|
+
if use_cache:
|
|
137
|
+
CACHE_DIR.mkdir(parents=True, exist_ok=True)
|
|
138
|
+
_cache_path(
|
|
139
|
+
job["text"], job["voice"], job.get("rate"), job.get("pitch")
|
|
140
|
+
).write_bytes(wav)
|
|
141
|
+
if progress:
|
|
142
|
+
progress(job, False)
|
|
143
|
+
|
|
144
|
+
return results
|
|
145
|
+
|
|
146
|
+
|
|
147
|
+
def list_voices(locale_prefix=None):
|
|
148
|
+
"""Liste les voix Edge TTS, filtrees par prefixe de locale (ex. 'fr')."""
|
|
149
|
+
_inject_truststore()
|
|
150
|
+
import edge_tts
|
|
151
|
+
|
|
152
|
+
voices = asyncio.run(edge_tts.list_voices())
|
|
153
|
+
if locale_prefix:
|
|
154
|
+
voices = [
|
|
155
|
+
v for v in voices
|
|
156
|
+
if v["Locale"].lower().startswith(locale_prefix.lower())
|
|
157
|
+
]
|
|
158
|
+
return sorted(voices, key=lambda v: v["ShortName"])
|
|
159
|
+
|
|
160
|
+
|
|
161
|
+
def wav_duration_ms(wav_bytes):
|
|
162
|
+
with wave.open(io.BytesIO(wav_bytes)) as w:
|
|
163
|
+
return int(round(1000.0 * w.getnframes() / w.getframerate()))
|