Az-r-ow
commited on
Commit
·
65b2047
1
Parent(s):
4e0aa6b
WIP(interface): added map with route
Browse files- app/app.py +95 -30
- test1.txt +1 -0
app/app.py
CHANGED
|
@@ -8,6 +8,7 @@ from travel_resolver.libs.nlp.ner.data_processing import process_sentence
|
|
| 8 |
from travel_resolver.libs.pathfinder.CSVTravelGraph import CSVTravelGraph
|
| 9 |
from travel_resolver.libs.pathfinder.graph import Graph
|
| 10 |
import time
|
|
|
|
| 11 |
|
| 12 |
transcriber = pipeline(
|
| 13 |
"automatic-speech-recognition", model="openai/whisper-base", device="cpu"
|
|
@@ -16,21 +17,14 @@ transcriber = pipeline(
|
|
| 16 |
models = {"LSTM": LSTM_NER(), "BiLSTM": BiLSTM_NER(), "CamemBERT": CamemBERT_NER()}
|
| 17 |
|
| 18 |
|
| 19 |
-
def handle_model_change(audio, file, model):
|
| 20 |
-
if audio:
|
| 21 |
-
render_tabs([transcribe(audio)], model, gr.Progress())
|
| 22 |
-
elif file:
|
| 23 |
-
with open(file.name, "r") as f:
|
| 24 |
-
sentences = f.read().split("\n")
|
| 25 |
-
return render_tabs(sentences, model, gr.Progress())
|
| 26 |
-
|
| 27 |
-
|
| 28 |
def handle_audio(audio, model, progress=gr.Progress()):
|
| 29 |
progress(
|
| 30 |
0,
|
| 31 |
)
|
| 32 |
promptAudio = transcribe(audio)
|
| 33 |
|
|
|
|
|
|
|
| 34 |
time.sleep(1)
|
| 35 |
|
| 36 |
return render_tabs([promptAudio], model, progress)
|
|
@@ -69,28 +63,56 @@ with gr.Blocks() as demo:
|
|
| 69 |
interactive=True,
|
| 70 |
)
|
| 71 |
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
audio
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
def handleCityChange(city):
|
| 81 |
-
stations = getStationsByCityName(city)
|
| 82 |
-
return gr.update(choices=stations, value=stations[0], interactive=True)
|
| 83 |
|
| 84 |
|
| 85 |
def handleCityChange(city):
|
| 86 |
stations = getStationsByCityName(city)
|
| 87 |
-
return gr.update(
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
|
| 89 |
|
| 90 |
def formatPath(path):
|
| 91 |
return "\n".join([f"{i + 1}. {elem}" for i, elem in enumerate(path)])
|
| 92 |
|
| 93 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
def handleStationChange(departureStation, destinationStation):
|
| 95 |
if departureStation and destinationStation:
|
| 96 |
dijkstraPath, dijkstraCost = getDijkstraResult(
|
|
@@ -98,18 +120,21 @@ def handleStationChange(departureStation, destinationStation):
|
|
| 98 |
)
|
| 99 |
dijkstraPathFormatted = formatPath(dijkstraPath)
|
| 100 |
AStarPath, AStarCost = getAStarResult(departureStation, destinationStation)
|
|
|
|
| 101 |
AStarPathFormatted = formatPath(AStarPath)
|
| 102 |
return (
|
| 103 |
gr.update(value=dijkstraCost),
|
| 104 |
gr.update(value=dijkstraPathFormatted, lines=len(dijkstraPath)),
|
| 105 |
gr.update(value=AStarCost),
|
| 106 |
gr.update(value=AStarPathFormatted, lines=len(AStarPath)),
|
|
|
|
| 107 |
)
|
| 108 |
return (
|
| 109 |
gr.HTML(HTML_COMPONENTS.NO_PROMPT.value),
|
| 110 |
gr.update(value=""),
|
| 111 |
gr.HTML(HTML_COMPONENTS.NO_PROMPT.value),
|
| 112 |
gr.update(value=""),
|
|
|
|
| 113 |
)
|
| 114 |
|
| 115 |
|
|
@@ -174,8 +199,24 @@ def getAStarResult(depart, destination):
|
|
| 174 |
|
| 175 |
def getStationsByCityName(city: str):
|
| 176 |
data = pd.read_csv("../data/sncf/gares_info.csv", sep=",")
|
| 177 |
-
stations =
|
| 178 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 179 |
|
| 180 |
|
| 181 |
def getEntitiesPositions(text, entity):
|
|
@@ -225,6 +266,7 @@ def render_tabs(sentences: list[str], model: str, progress_bar: gr.Progress):
|
|
| 225 |
for sentence in progress_bar.tqdm(sentences, desc=PROGRESS.PROCESSING.value):
|
| 226 |
with gr.Tab(f"Sentence {idx}"):
|
| 227 |
dep, arr = getDepartureAndArrivalFromText(sentence, model)
|
|
|
|
| 228 |
entities = []
|
| 229 |
for entity in [dep, arr]:
|
| 230 |
if entity:
|
|
@@ -237,10 +279,16 @@ def render_tabs(sentences: list[str], model: str, progress_bar: gr.Progress):
|
|
| 237 |
# Get the available stations
|
| 238 |
departureStations = getStationsByCityName(departureCityValue)
|
| 239 |
departureStationValue = (
|
| 240 |
-
departureStations[0]
|
|
|
|
|
|
|
| 241 |
)
|
| 242 |
arrivalStations = getStationsByCityName(arrivalCityValue)
|
| 243 |
-
arrivalStationValue =
|
|
|
|
|
|
|
|
|
|
|
|
|
| 244 |
|
| 245 |
dijkstraPathValues = []
|
| 246 |
AStarPathValues = []
|
|
@@ -255,6 +303,7 @@ def render_tabs(sentences: list[str], model: str, progress_bar: gr.Progress):
|
|
| 255 |
AStarPathValues, timeAStarValue = getAStarResult(
|
| 256 |
departureStationValue, arrivalStationValue
|
| 257 |
)
|
|
|
|
| 258 |
|
| 259 |
dijkstraPathFormatted = formatPath(dijkstraPathValues)
|
| 260 |
AStarPathFormatted = formatPath(AStarPathValues)
|
|
@@ -276,14 +325,19 @@ def render_tabs(sentences: list[str], model: str, progress_bar: gr.Progress):
|
|
| 276 |
with gr.Row():
|
| 277 |
departureStation = gr.Dropdown(
|
| 278 |
label="Gare de départ",
|
| 279 |
-
choices=departureStations,
|
| 280 |
value=departureStationValue,
|
| 281 |
)
|
| 282 |
arrivalStation = gr.Dropdown(
|
| 283 |
label="Gare d'arrivée",
|
| 284 |
-
choices=arrivalStations,
|
| 285 |
value=arrivalStationValue,
|
| 286 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 287 |
with gr.Tab("Dijkstra"):
|
| 288 |
timeDijkstra = gr.HTML(value=timeDijkstraValue)
|
| 289 |
dijkstraPath = gr.Textbox(
|
|
@@ -313,16 +367,27 @@ def render_tabs(sentences: list[str], model: str, progress_bar: gr.Progress):
|
|
| 313 |
departureStation.change(
|
| 314 |
handleStationChange,
|
| 315 |
inputs=[departureStation, arrivalStation],
|
| 316 |
-
outputs=[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 317 |
)
|
| 318 |
arrivalStation.change(
|
| 319 |
handleStationChange,
|
| 320 |
inputs=[departureStation, arrivalStation],
|
| 321 |
-
outputs=[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 322 |
)
|
| 323 |
|
| 324 |
idx += 1
|
| 325 |
-
return tabs
|
| 326 |
|
| 327 |
|
| 328 |
if __name__ == "__main__":
|
|
|
|
| 8 |
from travel_resolver.libs.pathfinder.CSVTravelGraph import CSVTravelGraph
|
| 9 |
from travel_resolver.libs.pathfinder.graph import Graph
|
| 10 |
import time
|
| 11 |
+
import plotly.graph_objects as go
|
| 12 |
|
| 13 |
transcriber = pipeline(
|
| 14 |
"automatic-speech-recognition", model="openai/whisper-base", device="cpu"
|
|
|
|
| 17 |
models = {"LSTM": LSTM_NER(), "BiLSTM": BiLSTM_NER(), "CamemBERT": CamemBERT_NER()}
|
| 18 |
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
def handle_audio(audio, model, progress=gr.Progress()):
|
| 21 |
progress(
|
| 22 |
0,
|
| 23 |
)
|
| 24 |
promptAudio = transcribe(audio)
|
| 25 |
|
| 26 |
+
print(f"prompt : {promptAudio}")
|
| 27 |
+
|
| 28 |
time.sleep(1)
|
| 29 |
|
| 30 |
return render_tabs([promptAudio], model, progress)
|
|
|
|
| 63 |
interactive=True,
|
| 64 |
)
|
| 65 |
|
| 66 |
+
@gr.render(
|
| 67 |
+
inputs=[audio, file, model], triggers=[audio.change, file.upload, model.change]
|
| 68 |
+
)
|
| 69 |
+
def handle_changes(audio, file, model):
|
| 70 |
+
if audio:
|
| 71 |
+
return handle_audio(audio, model)
|
| 72 |
+
elif file:
|
| 73 |
+
return handle_file(file, model)
|
|
|
|
|
|
|
|
|
|
| 74 |
|
| 75 |
|
| 76 |
def handleCityChange(city):
|
| 77 |
stations = getStationsByCityName(city)
|
| 78 |
+
return gr.update(
|
| 79 |
+
choices=[station["Nom de le gare"] for station in stations],
|
| 80 |
+
value=stations[0]["Nom de la gare"],
|
| 81 |
+
interactive=True,
|
| 82 |
+
)
|
| 83 |
|
| 84 |
|
| 85 |
def formatPath(path):
|
| 86 |
return "\n".join([f"{i + 1}. {elem}" for i, elem in enumerate(path)])
|
| 87 |
|
| 88 |
|
| 89 |
+
def plotMap(stationsInformation: dict):
|
| 90 |
+
stationNames = stationsInformation["stations"] if len(stationsInformation) else []
|
| 91 |
+
stationsLat = stationsInformation["lat"] if len(stationsInformation) else []
|
| 92 |
+
stationsLon = stationsInformation["lon"] if len(stationsInformation) else []
|
| 93 |
+
|
| 94 |
+
plt = go.Figure(
|
| 95 |
+
go.Scattermapbox(
|
| 96 |
+
lat=stationsLat,
|
| 97 |
+
lon=stationsLon,
|
| 98 |
+
mode="markers+lines",
|
| 99 |
+
marker=go.scattermapbox.Marker(size=14),
|
| 100 |
+
text=stationNames,
|
| 101 |
+
)
|
| 102 |
+
)
|
| 103 |
+
|
| 104 |
+
plt.update_layout(
|
| 105 |
+
mapbox_style="open-street-map",
|
| 106 |
+
mapbox=dict(
|
| 107 |
+
center=go.layout.mapbox.Center(lat=stationsLat[0], lon=stationsLon[0]),
|
| 108 |
+
pitch=0,
|
| 109 |
+
zoom=3,
|
| 110 |
+
),
|
| 111 |
+
)
|
| 112 |
+
|
| 113 |
+
return plt
|
| 114 |
+
|
| 115 |
+
|
| 116 |
def handleStationChange(departureStation, destinationStation):
|
| 117 |
if departureStation and destinationStation:
|
| 118 |
dijkstraPath, dijkstraCost = getDijkstraResult(
|
|
|
|
| 120 |
)
|
| 121 |
dijkstraPathFormatted = formatPath(dijkstraPath)
|
| 122 |
AStarPath, AStarCost = getAStarResult(departureStation, destinationStation)
|
| 123 |
+
AStarStationsInformation = getStationsInformation(AStarPath)
|
| 124 |
AStarPathFormatted = formatPath(AStarPath)
|
| 125 |
return (
|
| 126 |
gr.update(value=dijkstraCost),
|
| 127 |
gr.update(value=dijkstraPathFormatted, lines=len(dijkstraPath)),
|
| 128 |
gr.update(value=AStarCost),
|
| 129 |
gr.update(value=AStarPathFormatted, lines=len(AStarPath)),
|
| 130 |
+
plotMap(AStarStationsInformation),
|
| 131 |
)
|
| 132 |
return (
|
| 133 |
gr.HTML(HTML_COMPONENTS.NO_PROMPT.value),
|
| 134 |
gr.update(value=""),
|
| 135 |
gr.HTML(HTML_COMPONENTS.NO_PROMPT.value),
|
| 136 |
gr.update(value=""),
|
| 137 |
+
gr.update(value=plotMap(AStarStationsInformation)),
|
| 138 |
)
|
| 139 |
|
| 140 |
|
|
|
|
| 199 |
|
| 200 |
def getStationsByCityName(city: str):
|
| 201 |
data = pd.read_csv("../data/sncf/gares_info.csv", sep=",")
|
| 202 |
+
stations = data[data["Commune"] == city]
|
| 203 |
+
return dict(
|
| 204 |
+
stations=stations["Nom de la gare"].to_list(),
|
| 205 |
+
lat=stations["Latitude"].to_list(),
|
| 206 |
+
lon=stations["Longitude"].to_list(),
|
| 207 |
+
)
|
| 208 |
+
|
| 209 |
+
|
| 210 |
+
def getStationsInformation(stations: list[str]):
|
| 211 |
+
data = pd.read_csv("../data/sncf/gares_info.csv", sep=",")
|
| 212 |
+
data = data[data["Nom de la gare"].isin(stations)]
|
| 213 |
+
print(stations)
|
| 214 |
+
print(data)
|
| 215 |
+
return dict(
|
| 216 |
+
stations=data["Nom de la gare"].to_list(),
|
| 217 |
+
lat=data["Latitude"].to_list(),
|
| 218 |
+
lon=data["Longitude"].to_list(),
|
| 219 |
+
)
|
| 220 |
|
| 221 |
|
| 222 |
def getEntitiesPositions(text, entity):
|
|
|
|
| 266 |
for sentence in progress_bar.tqdm(sentences, desc=PROGRESS.PROCESSING.value):
|
| 267 |
with gr.Tab(f"Sentence {idx}"):
|
| 268 |
dep, arr = getDepartureAndArrivalFromText(sentence, model)
|
| 269 |
+
print(f"dep: {dep}, arr: {arr}")
|
| 270 |
entities = []
|
| 271 |
for entity in [dep, arr]:
|
| 272 |
if entity:
|
|
|
|
| 279 |
# Get the available stations
|
| 280 |
departureStations = getStationsByCityName(departureCityValue)
|
| 281 |
departureStationValue = (
|
| 282 |
+
departureStations["stations"][0]
|
| 283 |
+
if len(departureStations["stations"])
|
| 284 |
+
else ""
|
| 285 |
)
|
| 286 |
arrivalStations = getStationsByCityName(arrivalCityValue)
|
| 287 |
+
arrivalStationValue = (
|
| 288 |
+
arrivalStations["stations"][0]
|
| 289 |
+
if len(arrivalStations["stations"])
|
| 290 |
+
else ""
|
| 291 |
+
)
|
| 292 |
|
| 293 |
dijkstraPathValues = []
|
| 294 |
AStarPathValues = []
|
|
|
|
| 303 |
AStarPathValues, timeAStarValue = getAStarResult(
|
| 304 |
departureStationValue, arrivalStationValue
|
| 305 |
)
|
| 306 |
+
AStarStationsInformation = getStationsInformation(AStarPathValues)
|
| 307 |
|
| 308 |
dijkstraPathFormatted = formatPath(dijkstraPathValues)
|
| 309 |
AStarPathFormatted = formatPath(AStarPathValues)
|
|
|
|
| 325 |
with gr.Row():
|
| 326 |
departureStation = gr.Dropdown(
|
| 327 |
label="Gare de départ",
|
| 328 |
+
choices=departureStations["stations"],
|
| 329 |
value=departureStationValue,
|
| 330 |
)
|
| 331 |
arrivalStation = gr.Dropdown(
|
| 332 |
label="Gare d'arrivée",
|
| 333 |
+
choices=arrivalStations["stations"],
|
| 334 |
value=arrivalStationValue,
|
| 335 |
)
|
| 336 |
+
|
| 337 |
+
plt = plotMap(AStarStationsInformation)
|
| 338 |
+
|
| 339 |
+
map = gr.Plot(plt)
|
| 340 |
+
|
| 341 |
with gr.Tab("Dijkstra"):
|
| 342 |
timeDijkstra = gr.HTML(value=timeDijkstraValue)
|
| 343 |
dijkstraPath = gr.Textbox(
|
|
|
|
| 367 |
departureStation.change(
|
| 368 |
handleStationChange,
|
| 369 |
inputs=[departureStation, arrivalStation],
|
| 370 |
+
outputs=[
|
| 371 |
+
timeDijkstra,
|
| 372 |
+
dijkstraPath,
|
| 373 |
+
timeAStar,
|
| 374 |
+
AstarPath,
|
| 375 |
+
map,
|
| 376 |
+
],
|
| 377 |
)
|
| 378 |
arrivalStation.change(
|
| 379 |
handleStationChange,
|
| 380 |
inputs=[departureStation, arrivalStation],
|
| 381 |
+
outputs=[
|
| 382 |
+
timeDijkstra,
|
| 383 |
+
dijkstraPath,
|
| 384 |
+
timeAStar,
|
| 385 |
+
AstarPath,
|
| 386 |
+
map,
|
| 387 |
+
],
|
| 388 |
)
|
| 389 |
|
| 390 |
idx += 1
|
|
|
|
| 391 |
|
| 392 |
|
| 393 |
if __name__ == "__main__":
|
test1.txt
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
Je veux prendre le train de Lyon à Marseille.
|