Spaces:
Running
Running
Update src/streamlit_app.py
Browse files- src/streamlit_app.py +122 -67
src/streamlit_app.py
CHANGED
@@ -1,8 +1,13 @@
|
|
|
|
|
|
1 |
import streamlit as st
|
2 |
import io
|
3 |
import csv
|
4 |
-
|
5 |
from segments import SegmentsClient
|
|
|
|
|
|
|
6 |
from get_labels_from_samples import (
|
7 |
get_samples as get_samples_objects,
|
8 |
export_frames_and_annotations,
|
@@ -44,6 +49,95 @@ def parse_classes(input_str: str) -> list:
|
|
44 |
return sorted(set(classes))
|
45 |
|
46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
def generate_csv(metrics: list, dataset_identifier: str) -> str:
|
48 |
"""
|
49 |
Generate CSV content from list of per-sample metrics.
|
@@ -102,6 +196,9 @@ if api_key and dataset_identifier:
|
|
102 |
if is_multisensor:
|
103 |
sensor_select = st.selectbox("Choose sensor (optional)", options=['All sensors'] + sensor_names)
|
104 |
|
|
|
|
|
|
|
105 |
if run_button:
|
106 |
st.session_state.csv_content = None
|
107 |
st.session_state.error = None
|
@@ -122,75 +219,33 @@ if run_button:
|
|
122 |
st.info("Checking dataset type...")
|
123 |
try:
|
124 |
target_classes = parse_classes(classes_input)
|
125 |
-
|
126 |
metrics = []
|
127 |
# Update loader after dataset type check
|
128 |
if status_ctx is not None:
|
129 |
status_ctx.update(label="Dataset type checked. Processing samples...", state="running")
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
-
|
134 |
-
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
-
matching_annotations = sum(
|
153 |
-
1
|
154 |
-
for f in frames_list
|
155 |
-
for ann in f['annotations']
|
156 |
-
if getattr(ann, 'category_id', None) in target_classes
|
157 |
-
)
|
158 |
-
metrics.append({
|
159 |
-
'name': getattr(sample, 'name', sample.uuid),
|
160 |
-
'uuid': sample.uuid,
|
161 |
-
'labelset': labelset,
|
162 |
-
'sensor': sensor_name,
|
163 |
-
'num_frames': num_frames,
|
164 |
-
'total_annotations': total_annotations,
|
165 |
-
'matching_annotations': matching_annotations,
|
166 |
-
'labeled_by': labeled_by,
|
167 |
-
'reviewed_by': reviewed_by
|
168 |
-
})
|
169 |
-
continue
|
170 |
-
else:
|
171 |
-
frames_list = export_frames_and_annotations(label)
|
172 |
-
sensor_val = ''
|
173 |
-
num_frames = len(frames_list)
|
174 |
-
total_annotations = sum(len(f['annotations']) for f in frames_list)
|
175 |
-
matching_annotations = sum(
|
176 |
-
1
|
177 |
-
for f in frames_list
|
178 |
-
for ann in f['annotations']
|
179 |
-
if getattr(ann, 'category_id', None) in target_classes
|
180 |
-
)
|
181 |
-
metrics.append({
|
182 |
-
'name': getattr(sample, 'name', sample.uuid),
|
183 |
-
'uuid': sample.uuid,
|
184 |
-
'labelset': labelset,
|
185 |
-
'sensor': sensor_val if is_multisensor else '',
|
186 |
-
'num_frames': num_frames,
|
187 |
-
'total_annotations': total_annotations,
|
188 |
-
'matching_annotations': matching_annotations,
|
189 |
-
'labeled_by': labeled_by,
|
190 |
-
'reviewed_by': reviewed_by
|
191 |
-
})
|
192 |
-
except Exception as e:
|
193 |
-
continue
|
194 |
if not metrics:
|
195 |
st.session_state.error = "No metrics could be generated for the dataset."
|
196 |
else:
|
@@ -213,4 +268,4 @@ if st.session_state.csv_content:
|
|
213 |
data=st.session_state.csv_content,
|
214 |
file_name=filename,
|
215 |
mime="text/csv"
|
216 |
-
)
|
|
|
1 |
+
#!/usr/bin/env python3
|
2 |
+
|
3 |
import streamlit as st
|
4 |
import io
|
5 |
import csv
|
6 |
+
import concurrent.futures
|
7 |
from segments import SegmentsClient
|
8 |
+
from datetime import datetime
|
9 |
+
import sys
|
10 |
+
import os
|
11 |
from get_labels_from_samples import (
|
12 |
get_samples as get_samples_objects,
|
13 |
export_frames_and_annotations,
|
|
|
49 |
return sorted(set(classes))
|
50 |
|
51 |
|
52 |
+
def _count_from_frames(frames, target_set):
|
53 |
+
"""Helper to count frames, total annotations, and matching annotations directly."""
|
54 |
+
if not frames:
|
55 |
+
return 0, 0, 0
|
56 |
+
num_frames = len(frames)
|
57 |
+
total_annotations = 0
|
58 |
+
matching_annotations = 0
|
59 |
+
for f in frames:
|
60 |
+
anns = getattr(f, 'annotations', [])
|
61 |
+
total_annotations += len(anns)
|
62 |
+
if target_set:
|
63 |
+
for ann in anns:
|
64 |
+
if getattr(ann, 'category_id', None) in target_set:
|
65 |
+
matching_annotations += 1
|
66 |
+
return num_frames, total_annotations, matching_annotations
|
67 |
+
|
68 |
+
|
69 |
+
def compute_metrics_for_sample(sample, api_key, target_set, is_multisensor, sensor_select):
|
70 |
+
"""
|
71 |
+
Fetch label for a single sample and compute metrics.
|
72 |
+
Returns a list of metric dicts (one per sensor if 'All sensors', otherwise one).
|
73 |
+
"""
|
74 |
+
try:
|
75 |
+
client = init_client(api_key)
|
76 |
+
label = client.get_label(sample.uuid)
|
77 |
+
labelset = getattr(label, 'labelset', '') or ''
|
78 |
+
labeled_by = getattr(label, 'created_by', '') or ''
|
79 |
+
reviewed_by = getattr(label, 'reviewed_by', '') or ''
|
80 |
+
|
81 |
+
metrics_rows = []
|
82 |
+
|
83 |
+
if is_multisensor:
|
84 |
+
sensors = getattr(getattr(label, 'attributes', None), 'sensors', None) or []
|
85 |
+
if sensor_select and sensor_select != 'All sensors':
|
86 |
+
# single sensor
|
87 |
+
for sensor in sensors:
|
88 |
+
if getattr(sensor, 'name', None) == sensor_select:
|
89 |
+
frames = getattr(getattr(sensor, 'attributes', None), 'frames', [])
|
90 |
+
num_frames, total_annotations, matching_annotations = _count_from_frames(frames, target_set)
|
91 |
+
metrics_rows.append({
|
92 |
+
'name': getattr(sample, 'name', sample.uuid),
|
93 |
+
'uuid': sample.uuid,
|
94 |
+
'labelset': labelset,
|
95 |
+
'sensor': sensor_select,
|
96 |
+
'num_frames': num_frames,
|
97 |
+
'total_annotations': total_annotations,
|
98 |
+
'matching_annotations': matching_annotations,
|
99 |
+
'labeled_by': labeled_by,
|
100 |
+
'reviewed_by': reviewed_by
|
101 |
+
})
|
102 |
+
break
|
103 |
+
else:
|
104 |
+
# all sensors
|
105 |
+
for sensor in sensors:
|
106 |
+
sensor_name = getattr(sensor, 'name', 'Unknown')
|
107 |
+
frames = getattr(getattr(sensor, 'attributes', None), 'frames', [])
|
108 |
+
num_frames, total_annotations, matching_annotations = _count_from_frames(frames, target_set)
|
109 |
+
metrics_rows.append({
|
110 |
+
'name': getattr(sample, 'name', sample.uuid),
|
111 |
+
'uuid': sample.uuid,
|
112 |
+
'labelset': labelset,
|
113 |
+
'sensor': sensor_name,
|
114 |
+
'num_frames': num_frames,
|
115 |
+
'total_annotations': total_annotations,
|
116 |
+
'matching_annotations': matching_annotations,
|
117 |
+
'labeled_by': labeled_by,
|
118 |
+
'reviewed_by': reviewed_by
|
119 |
+
})
|
120 |
+
else:
|
121 |
+
# single-sensor dataset
|
122 |
+
frames = getattr(getattr(label, 'attributes', None), 'frames', [])
|
123 |
+
num_frames, total_annotations, matching_annotations = _count_from_frames(frames, target_set)
|
124 |
+
metrics_rows.append({
|
125 |
+
'name': getattr(sample, 'name', sample.uuid),
|
126 |
+
'uuid': sample.uuid,
|
127 |
+
'labelset': labelset,
|
128 |
+
'sensor': '',
|
129 |
+
'num_frames': num_frames,
|
130 |
+
'total_annotations': total_annotations,
|
131 |
+
'matching_annotations': matching_annotations,
|
132 |
+
'labeled_by': labeled_by,
|
133 |
+
'reviewed_by': reviewed_by
|
134 |
+
})
|
135 |
+
|
136 |
+
return metrics_rows
|
137 |
+
except Exception:
|
138 |
+
return []
|
139 |
+
|
140 |
+
|
141 |
def generate_csv(metrics: list, dataset_identifier: str) -> str:
|
142 |
"""
|
143 |
Generate CSV content from list of per-sample metrics.
|
|
|
196 |
if is_multisensor:
|
197 |
sensor_select = st.selectbox("Choose sensor (optional)", options=['All sensors'] + sensor_names)
|
198 |
|
199 |
+
# Concurrency control
|
200 |
+
parallel_workers = st.slider("Parallel requests", min_value=1, max_value=32, value=8, help="Increase to speed up processing; lower if you hit API limits.")
|
201 |
+
|
202 |
if run_button:
|
203 |
st.session_state.csv_content = None
|
204 |
st.session_state.error = None
|
|
|
219 |
st.info("Checking dataset type...")
|
220 |
try:
|
221 |
target_classes = parse_classes(classes_input)
|
222 |
+
target_set = set(target_classes)
|
223 |
metrics = []
|
224 |
# Update loader after dataset type check
|
225 |
if status_ctx is not None:
|
226 |
status_ctx.update(label="Dataset type checked. Processing samples...", state="running")
|
227 |
+
progress = st.progress(0)
|
228 |
+
total = len(samples_objects)
|
229 |
+
done = 0
|
230 |
+
with concurrent.futures.ThreadPoolExecutor(max_workers=parallel_workers) as executor:
|
231 |
+
futures = [
|
232 |
+
executor.submit(
|
233 |
+
compute_metrics_for_sample,
|
234 |
+
sample,
|
235 |
+
api_key,
|
236 |
+
target_set,
|
237 |
+
is_multisensor,
|
238 |
+
sensor_select,
|
239 |
+
)
|
240 |
+
for sample in samples_objects
|
241 |
+
]
|
242 |
+
for future in concurrent.futures.as_completed(futures):
|
243 |
+
rows = future.result()
|
244 |
+
if rows:
|
245 |
+
metrics.extend(rows)
|
246 |
+
done += 1
|
247 |
+
if total:
|
248 |
+
progress.progress(min(done / total, 1.0))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
249 |
if not metrics:
|
250 |
st.session_state.error = "No metrics could be generated for the dataset."
|
251 |
else:
|
|
|
268 |
data=st.session_state.csv_content,
|
269 |
file_name=filename,
|
270 |
mime="text/csv"
|
271 |
+
)
|