File size: 8,068 Bytes
0d27c43
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
"""
File: app.py
Author: Dr. Gordon Wright
Description: Six-emotion replication challenge. The user poses each of
             the six basic emotions in turn; the classifier reads each
             attempt; the tiles fill in to form a single-page A4
             EmotionMap artefact at the end.

             Compact layout: one-line header, collapsible privacy,
             single input row, 2x3 tile grid, one download button.
             Wireframe toggle anonymises every face on screen and in
             the export.
License: MIT License
"""

import gradio as gr
from PIL import Image as PILImage

from app.app_utils import preprocess_image_and_predict
from app.session import (
    BASIC_EMOTIONS,
    Capture,
    empty_session,
    session_status,
)
from app.tiles import grid_tiles, export_single_page_pdf


HEADER_HTML = """
<div class="te-header">
  <span class="te-title">Six-emotion replication challenge</span>
  <span class="te-sub">pose each emotion · classifier reads what it sees · download a one-page EmotionMap</span>
</div>
<details class="te-privacy">
  <summary>Privacy</summary>
  <p>This app uploads nothing beyond the model inference call and stores
  nothing on a server. Your captures live in this browser tab only — refresh
  and they are gone. Turn on <strong>Wireframe (face-free)</strong> below if
  you would rather not submit identifiable photos; the export then ships only
  the anonymised landmark mesh.</p>
</details>
"""

FOOTER_HTML = """
<div class="te-footer">
  Created by Dr. Gordon Wright — A LittleMonkeyLab caper.
  Part of the Goldsmiths MSc in Psychology, Week 3 Part 4.
</div>
"""


def _tiles_and_status(state, wireframe):
    image_size = None
    for cap in state.values():
        if cap is not None and cap.image_size is not None:
            image_size = cap.image_size
            break
    return grid_tiles(state, wireframe=wireframe, image_size=image_size), session_status(state)


def submit_attempt(image, intended, wireframe, state):
    if state is None:
        state = empty_session()

    if image is None:
        tiles, status = _tiles_and_status(state, wireframe)
        return state, *tiles, status + "  (Upload or capture a face first.)"

    if intended not in BASIC_EMOTIONS:
        tiles, status = _tiles_and_status(state, wireframe)
        return state, *tiles, status + "  (Pick which emotion you are posing.)"

    image_size = image.size  # (W, H)
    face, heatmap, confidences, blendshapes, landmarks, bbox = (
        preprocess_image_and_predict(image)
    )
    if face is None:
        tiles, status = _tiles_and_status(state, wireframe)
        return state, *tiles, status + "  (No face detected in that image.)"

    state[intended] = Capture(
        intended=intended,
        face=face,
        emotion_probs=confidences,
        heatmap=heatmap,
        blendshapes=blendshapes,
        landmarks=landmarks,
        bbox=bbox,
        image_size=image_size,
    )
    tiles, status = _tiles_and_status(state, wireframe)
    return state, *tiles, status


def retry_slot(slot_emotion, wireframe, state):
    if state is None:
        state = empty_session()
    state[slot_emotion] = None
    tiles, status = _tiles_and_status(state, wireframe)
    return state, *tiles, status


def clear_all(_state, wireframe):
    state = empty_session()
    tiles, status = _tiles_and_status(state, wireframe)
    return state, *tiles, status


def toggle_wireframe(wireframe, state):
    tiles, status = _tiles_and_status(state or empty_session(), wireframe)
    return tiles + [status]


def download_artefact(state, wireframe, student_name):
    if state is None:
        state = empty_session()
    image_size = None
    for cap in state.values():
        if cap is not None and cap.image_size is not None:
            image_size = cap.image_size
            break
    return export_single_page_pdf(
        state,
        student_name=student_name or "",
        wireframe=wireframe,
        image_size=image_size,
    )


with gr.Blocks(title="totes-emosh") as demo:
    gr.HTML(HEADER_HTML)

    session_state = gr.State(empty_session())

    gr.HTML(
        '<div class="te-steps">'
        '<span class="te-step"><b>1.</b> Take a photo with your webcam, '
        'or drop in an image</span>'
        '<span class="te-step"><b>2.</b> Pick which of the six emotions '
        'you are posing</span>'
        '<span class="te-step"><b>3.</b> Submit — the matching tile '
        'fills in</span>'
        '</div>'
    )

    with gr.Row(elem_classes="te-input-row"):
        with gr.Column(scale=2):
            input_image = gr.Image(
                label="Your pose",
                type="pil",
                sources=["webcam", "upload"],
                height=320,
                elem_classes="te-input-image",
            )
        with gr.Column(scale=1):
            intended_emotion = gr.Dropdown(
                choices=BASIC_EMOTIONS,
                value="happy",
                label="Which emotion are you posing?",
            )
            wireframe_toggle = gr.Checkbox(
                value=False,
                label="Wireframe (face-free) mode",
            )
            submit_btn = gr.Button(
                value="Submit attempt",
                variant="primary",
                size="lg",
                elem_classes="te-submit",
            )
            clear_btn = gr.Button(value="Clear all six")

    status_md = gr.Markdown(value=session_status(empty_session()),
                            elem_classes="te-status")

    # 2x3 tile grid
    initial_tiles = grid_tiles(empty_session())
    tile_images = []
    retry_buttons = []
    for row in range(2):
        with gr.Row(elem_classes="te-tile-row"):
            for col in range(3):
                idx = row * 3 + col
                emo = BASIC_EMOTIONS[idx]
                with gr.Column(elem_classes="te-tile-col"):
                    img = gr.Image(
                        value=initial_tiles[idx],
                        show_label=False,
                        interactive=False,
                        height=406,
                        elem_classes="te-tile-img",
                    )
                    retry = gr.Button(
                        value=f"Retry {emo}",
                        size="sm",
                        elem_classes="te-retry",
                    )
                    tile_images.append(img)
                    retry_buttons.append((emo, retry))

    with gr.Row(elem_classes="te-export-row"):
        student_name = gr.Textbox(
            label="Your name (appears on the PDF)",
            placeholder="optional",
            scale=2,
        )
        download_btn = gr.Button(
            value="Download single-page EmotionMap PDF",
            variant="primary",
            scale=1,
        )
    download_file = gr.File(label="EmotionMap PDF", interactive=False)

    gr.HTML(FOOTER_HTML)

    # ---- wiring ----
    submit_outputs = [session_state, *tile_images, status_md]
    submit_btn.click(
        fn=submit_attempt,
        inputs=[input_image, intended_emotion, wireframe_toggle, session_state],
        outputs=submit_outputs,
        queue=True,
    )

    clear_btn.click(
        fn=clear_all,
        inputs=[session_state, wireframe_toggle],
        outputs=submit_outputs,
        queue=False,
    )

    for slot_emo, button in retry_buttons:
        button.click(
            fn=retry_slot,
            inputs=[gr.State(slot_emo), wireframe_toggle, session_state],
            outputs=submit_outputs,
            queue=False,
        )

    wireframe_toggle.change(
        fn=toggle_wireframe,
        inputs=[wireframe_toggle, session_state],
        outputs=[*tile_images, status_md],
        queue=False,
    )

    download_btn.click(
        fn=download_artefact,
        inputs=[session_state, wireframe_toggle, student_name],
        outputs=[download_file],
        queue=True,
    )


if __name__ == "__main__":
    demo.queue(api_open=False).launch(share=False, css_paths=["app.css"])