Spaces:
Running
Running
File size: 14,584 Bytes
3daca56 029f30f 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a c08865b 952fa4a 3daca56 952fa4a 3daca56 c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 3daca56 952fa4a c08865b 952fa4a 3daca56 952fa4a 3daca56 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a 3daca56 952fa4a 3daca56 029f30f 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 029f30f 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 3daca56 952fa4a 029f30f 952fa4a 029f30f 952fa4a 3daca56 952fa4a 3daca56 952fa4a c08865b 3daca56 952fa4a 3daca56 952fa4a c08865b 952fa4a c08865b 952fa4a c08865b 952fa4a 029f30f 952fa4a 029f30f 952fa4a 029f30f 3daca56 952fa4a |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 |
import gradio as gr
import json
import os
import tempfile
LANG_DIR = "./human_eval"
SAVE_DIR = "./annotations"
os.makedirs(SAVE_DIR, exist_ok=True)
# 全局变量 data, user_annotations, current_lang 已被移除
language_options = sorted([f for f in os.listdir(LANG_DIR)])
# --- 函数修改 ---
# 每个需要访问或修改 data, user_annotations, current_lang 的函数
# 都需要将 app_state 作为输入,并通常作为输出
def load_data_for_lang(lang_pair, current_app_state): # 接收 app_state
file_path = os.path.join(LANG_DIR, lang_pair, f"{lang_pair}.json")
with open(file_path, "r", encoding="utf-8") as f:
loaded_data = json.load(f)
new_app_state = { # 创建新的状态
"data": loaded_data,
"user_annotations": [], # Reset annotations for new language
"current_lang": lang_pair
}
if not loaded_data:
return 0, "", "", f"0/0 loaded from {lang_pair}", new_app_state
return (
0,
loaded_data[0]["source"],
loaded_data[0]["hypothesis"],
f"0/{len(loaded_data)} loaded from {lang_pair}",
new_app_state, # 返回更新后的状态
)
def restore_previous_annotations(file_obj, current_app_state): # 接收 app_state
if file_obj is None: # Check if a file was actually uploaded
return 0, "", "", "No file uploaded.", current_app_state, language_options[0] if language_options else ""
with open(file_obj.name, "r", encoding="utf-8") as f:
uploaded_annotations = json.load(f)
if not uploaded_annotations:
return 0, "", "", "No annotations found in file.", current_app_state, language_options[0] if language_options else ""
restored_lang = uploaded_annotations[0].get("lang_pair", None)
if not restored_lang or not os.path.exists(
os.path.join(LANG_DIR, restored_lang, f"{restored_lang}.json")
):
return 0, "", "", "❌ Language pair info missing or file not found.", current_app_state, language_options[0] if language_options else ""
file_path = os.path.join(LANG_DIR, restored_lang, f"{restored_lang}.json")
with open(file_path, "r", encoding="utf-8") as f:
loaded_data = json.load(f)
new_app_state = { # 创建新的状态
"data": loaded_data,
"user_annotations": uploaded_annotations,
"current_lang": restored_lang
}
last_index = 0
if new_app_state["user_annotations"]: # Check if there are any annotations
last_index = new_app_state["user_annotations"][-1].get("index", -1) + 1
if last_index >= len(new_app_state["data"]):
return (
last_index,
"",
"",
f"✅ Already completed {len(new_app_state['data'])} samples of {restored_lang}.",
new_app_state, # 返回状态
restored_lang,
)
return (
last_index,
new_app_state["data"][last_index]["source"],
new_app_state["data"][last_index]["hypothesis"],
f"Restored {restored_lang}: {last_index}/{len(new_app_state['data'])}",
new_app_state, # 返回状态
restored_lang, # To update the dropdown
)
def load_sample(i, current_app_state): # 接收 app_state
# 这个函数只读取状态,不修改,所以不需要返回 app_state
# 但仍然需要接收它来获取数据
data_from_state = current_app_state["data"]
if not data_from_state or int(i) >= len(data_from_state) or int(i) < 0 :
return "", ""
entry = data_from_state[int(i)]
return entry["source"], entry["hypothesis"]
def annotate(index, score, comment, annotator, current_app_state): # 接收 app_state
index = int(index)
app_data = current_app_state["data"]
app_current_lang = current_app_state["current_lang"]
app_user_annotations = list(current_app_state["user_annotations"]) # Create a mutable copy
if index >= len(app_data): # Safety check
return (
"Error: Index out of bounds.",
index,
f"Error annotating.",
gr.update(),
gr.update(),
gr.update(),
gr.update(),
gr.update(visible=False),
current_app_state # Return original state
)
entry = app_data[index]
record = {
"index": index,
"annotator": annotator,
"lang_pair": app_current_lang,
"source": entry["source"],
"hypothesis": entry["hypothesis"],
"score": score,
"comment": comment,
}
# Update user_annotations in the copied list
app_user_annotations = [
rec
for rec in app_user_annotations
if not (rec["index"] == index and rec["annotator"] == annotator) # More robust removal
]
app_user_annotations.append(record)
app_user_annotations.sort(key=lambda x: x["index"]) # Keep sorted if needed
# 更新状态
new_app_state = {
"data": app_data,
"user_annotations": app_user_annotations,
"current_lang": app_current_lang
}
completed = index + 1
if completed >= len(app_data):
return (
"🎉 All samples annotated!",
index, # or completed
f"✅ Completed {completed}/{len(app_data)} samples.",
gr.update(interactive=False),
gr.update(interactive=False),
gr.update(interactive=False),
gr.update(interactive=False),
gr.update(visible=True),
new_app_state, # 返回状态
)
next_index = index + 1
next_entry = app_data[next_index]
prev_score, prev_comment = 0, ""
for rec in new_app_state["user_annotations"]:
if rec["index"] == next_index and rec["annotator"] == annotator:
prev_score = rec["score"]
prev_comment = rec["comment"]
break
progress_text = f"{completed}/{len(app_data)} annotated by {annotator}"
is_at_start = next_index == 0
return (
"✅ Saved",
next_index,
progress_text,
gr.update(value=prev_score, interactive=True),
gr.update(value=prev_comment, interactive=True),
gr.update(interactive=not is_at_start),
gr.update(interactive=True),
gr.update(visible=False),
new_app_state, # 返回状态
)
def go_previous(index, annotator, current_app_state): # 接收 app_state
index = int(index)
app_data = current_app_state["data"]
app_user_annotations = current_app_state["user_annotations"]
if not app_data: # No data loaded
return 0, "", "", 0, "", "No data loaded.", gr.update(interactive=False), gr.update(interactive=False)
if index <= 0:
prev_index = 0
is_at_start = True
else:
prev_index = index - 1
is_at_start = prev_index == 0
entry = app_data[prev_index]
prev_score, prev_comment = 0, ""
for rec in app_user_annotations:
if rec["index"] == prev_index and rec["annotator"] == annotator:
prev_score = rec["score"]
prev_comment = rec["comment"]
break
progress_text = f"{prev_index}/{len(app_data)} annotated by {annotator}"
if not app_data:
progress_text = "No data loaded."
# This function doesn't change the state, so no need to return current_app_state unless it was modified
# However, to be consistent with other functions that MIGHT modify state or if you plan to,
# it's good practice to include it. In this specific case, it's only reading.
return (
prev_index,
entry["source"],
entry["hypothesis"],
prev_score,
prev_comment,
progress_text,
gr.update(interactive=not is_at_start),
gr.update(interactive=True),
# No app_state in outputs here if it's not being changed
)
def export_results(current_app_state): # 接收 app_state
app_user_annotations = current_app_state["user_annotations"]
if not app_user_annotations:
# raise ValueError("No annotations to export.") # This will crash the app
# Instead, show a message or disable the button if no annotations
gr.Warning("No annotations to export.")
return None, gr.update(visible=False)
# Create a temporary file for download
with tempfile.NamedTemporaryFile(
delete=False, suffix=".json", mode="w", encoding="utf-8"
) as tmp:
json.dump(app_user_annotations, tmp, ensure_ascii=False, indent=2)
tmp_path = tmp.name # Get the path before closing
# The file is closed when exiting the 'with' block
return tmp_path, gr.update(visible=True, value=tmp_path)
# ======== UI ========
with gr.Blocks() as demo:
# Define session state
app_state = gr.State(
value={"data": [], "user_annotations": [], "current_lang": ""}
)
gr.Markdown("## 📝 Direct Assessment Annotation Tool")
with gr.Row():
lang_choice = gr.Dropdown(
label="Choose Language Pair",
choices=language_options,
value=language_options[0] if language_options else None, # Handle empty options
)
load_button = gr.Button("🔄 Load Data")
with gr.Row():
upload_file = gr.File(
label="📤 Upload Previous Annotations", file_types=[".json"]
)
export_button = gr.Button("📥 Export My Results")
with gr.Row():
annotator = gr.Textbox(
label="Annotator ID",
placeholder="Enter your name or ID",
value="annotator_1",
)
progress = gr.Textbox(label="Progress", interactive=False)
idx = gr.Number(value=0, visible=False, label="Current Index") # Added label for clarity if made visible
source = gr.Textbox(label="Source Sentence", interactive=False, lines=3)
hyp = gr.Textbox(label="Machine Translation", interactive=False, lines=3)
score = gr.Slider(0, 100, step=1, label="Translation Quality Score", value=0)
comment = gr.Textbox(lines=2, placeholder="Optional comment...", label="Comment")
output = gr.Textbox(label="Status", interactive=False)
previous_button = gr.Button("⏪Previous", interactive=False) # Initially disabled
next_button = gr.Button("⏩Next", interactive=False) # Initially disabled
export_file = gr.File(label="Download your results", visible=False, interactive=False)
# --- Component Event Handlers ---
# Add app_state to inputs and outputs where necessary
load_button.click(
fn=load_data_for_lang,
inputs=[lang_choice, app_state], # Add app_state
outputs=[idx, source, hyp, progress, app_state], # Add app_state
).then(
lambda: (gr.update(interactive=True), gr.update(interactive=False)), # Enable Next, Disable Prev
outputs=[next_button, previous_button]
)
upload_file.change(
fn=restore_previous_annotations,
inputs=[upload_file, app_state], # Add app_state
outputs=[idx, source, hyp, progress, app_state, lang_choice], # Add app_state and lang_choice to update dropdown
).then(
lambda x: (gr.update(interactive=True), gr.update(interactive=x!=0)), # Enable Next, Prev based on index
inputs=[idx],
outputs=[next_button, previous_button]
)
next_button.click(
fn=annotate,
inputs=[idx, score, comment, annotator, app_state], # Add app_state
outputs=[
output,
idx,
progress,
score, # To reset/update score for next item
comment, # To reset/update comment for next item
previous_button,
next_button,
export_file, # This is the download component
app_state, # Add app_state
],
)
previous_button.click(
fn=go_previous,
inputs=[idx, annotator, app_state], # Add app_state
outputs=[
idx,
source,
hyp,
score,
comment,
progress,
previous_button,
next_button,
# No app_state output if go_previous doesn't modify it
],
)
export_button.click(
fn=export_results,
inputs=[app_state], # Add app_state
outputs=[
export_file, # For the file content/path
export_file, # For updating visibility/value
],
)
# This loads the sample when the index changes, e.g., after load_data or annotate
idx.change(fn=load_sample, inputs=[idx, app_state], outputs=[source, hyp])
# Initial load logic (optional, if you want something on app start)
# Consider what should happen if language_options is empty
def initial_load_or_message(current_app_state):
if language_options:
# This duplicates load_data_for_lang slightly, could be refactored
lang_pair_to_load = language_options[0]
idx_val, src_val, hyp_val, prog_val, new_app_state = load_data_for_lang(lang_pair_to_load, current_app_state)
return idx_val, src_val, hyp_val, prog_val, new_app_state, lang_pair_to_load, gr.update(interactive=True), gr.update(interactive=False)
else:
return 0, "No languages found in LANG_DIR.", "", "Please add language files.", current_app_state, None, gr.update(interactive=False), gr.update(interactive=False)
# demo.load is tricky with state management if the function also needs to return the state.
# It's often better to trigger an initial load via a button or specific logic after UI setup.
# If you must use demo.load and it needs to initialize state, ensure it correctly returns the state.
# For simplicity, let's ensure load_button handles enabling next/prev:
# demo.load(
# fn=initial_load_or_message,
# inputs=[app_state],
# outputs=[idx, source, hyp, progress, app_state, lang_choice, next_button, previous_button]
# )
# A simpler demo.load without initial data loading:
demo.load(lambda: (gr.update(interactive=False), gr.update(interactive=False)), outputs=[next_button, previous_button])
if __name__ == "__main__":
demo.launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860))) |