Files
visual_critical_discourse_a…/src/web/app.py
T
2024-02-25 00:04:01 +01:00

152 lines
4.5 KiB
Python

from dash import Dash, Input, Output, State, ALL
import dash_bootstrap_components as dbc
from dash_auth import BasicAuth
import logging
from typing import List
from pydantic import ValidationError
import os
from .layout import app_layout
from src.database import (
connect,
get_visual_communication,
NoDocumentFoundException,
upsert_annotations,
ModelOutputs
)
# setup app
app = Dash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])
app.title = "visual critical discourse analysis".title()
app.layout = app_layout
server = app.server
# setup authentication
AUTH_DICT = {
os.getenv("DASH_AUTH_USERNAME"): os.getenv("DASH_AUTH_PASSWORD")
}
BasicAuth(app, AUTH_DICT)
# connect to database
collection, db, client = connect()
# define callbacks
@app.callback(
Output("alert-element", "is_open"),
Output("alert-element", "children"),
Input("alert-message", "data")
)
def show_alert(
msg: str | None
):
if msg is None or msg == "":
return False, ""
logging.info(f"updated alert message: {msg}")
return True, msg
@app.callback(
Output("alert-message", "data"),
Output("vis-com-name", "data"),
Output("image-container", "src"),
Output({"type": "annotation", "index": ALL}, "value"),
Input("next-button", "n_clicks"),
State("vis-com-name", "data"),
State("image-container", "src"),
State({"type": "annotation", "index": ALL}, "id"),
State({"type": "annotation", "index": ALL}, "value"),
prevent_initial_call=True,
)
def cycle_visual_communication_data(
n_clicks: int,
vis_com_name: str,
image_src: str,
annotation_keys: List,
annotation_values: List,
):
logging.info("began cycling visual communication data")
global collection
# prepare default response
response = [
"",
vis_com_name,
image_src,
annotation_values
]
# check if next-button clicked
if n_clicks == 0:
logging.info("stopping early: next-button has not yet been clicked")
return response
# check if visual communication name is set
if len(vis_com_name) > 0:
logging.info("saving annotations to database: %s", vis_com_name)
try:
# extract option keys
annotation_keys = [
elem["index"]
for elem in annotation_keys
]
# ensure all options are set
logging.info(annotation_keys)
for option, value in zip(annotation_keys, annotation_values):
if value is None:
raise ValueError(f"{option} is not set")
# prepare data to save
annotation_keys = [
elem.replace(' ', '_')
for elem
in annotation_keys
]
annotation_values = [
elem.replace(' ', '_').lower()
for elem
in annotation_values
]
annotation_map = {
key: value
for key, value
in zip(annotation_keys, annotation_values)
}
# instantiate ModelOutputs object
annotations = ModelOutputs.from_annotations(**annotation_map)
# save data to database
upsert_annotations(
collection=collection,
vis_com_name=vis_com_name,
annotations=annotations
)
except (ValueError, ValidationError) as exc:
msg = f"failed saving annotation: {exc}"
logging.warning(msg)
response[0] = msg
return tuple(response)
# get new visual communication
logging.info("trying to get new visual communication")
try:
# get data
vis_com = get_visual_communication(
collection=collection,
with_annotation=False
)
# set variables
vis_com_name = vis_com.name
image_src = vis_com.webencoded_image()
if vis_com.prediction is not None:
# TODO: update to use optional predictions
pass
else:
# reset annotations
annotation_values = [None for elem in annotation_values]
except NoDocumentFoundException:
msg = "no unannotated data in database"
logging.warning(msg)
response[0] = msg
return tuple(response)
else:
response[1] = vis_com_name
response[2] = image_src
response[3] = annotation_values
logging.info("finished getting visual communication: %s", vis_com_name)
return tuple(response)