Spaces:
Running
Running
File size: 13,012 Bytes
09801ca | 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 | # =============================================================================
# CHATGPT-STYLE CHART GENERATION - From Answer Data
# =============================================================================
#
# This module extracts chart data directly from LLM answers to ensure
# charts match the text response exactly - just like real ChatGPT!
#
from typing import Optional, Dict, List, Any, Tuple
import re
import json
def extract_chart_data_from_answer(
answer: str,
query: str,
df=None,
currency_symbol: str = "₹"
) -> Optional[Dict]:
"""
🧠 CHATGPT-STYLE: Extract chart data from LLM's text answer.
Instead of generating charts separately, we extract the data
that the LLM mentioned in its answer and visualize THAT.
Args:
answer: The LLM's text response
query: User's original query
df: Optional DataFrame for validation
currency_symbol: Currency symbol used
Returns:
Chart specification dict or None
"""
# Detect what kind of chart is needed
query_lower = query.lower()
# ===========================================================================
# STEP 1: Detect chart type from query
# ===========================================================================
chart_type = "bar" # Default
if "pie" in query_lower or "distribution" in query_lower or "share" in query_lower:
chart_type = "pie"
elif "line" in query_lower or "trend" in query_lower or "over time" in query_lower:
chart_type = "line"
elif "bar" in query_lower or "top" in query_lower or "ranking" in query_lower:
chart_type = "bar"
# ===========================================================================
# STEP 2: Extract structured data from answer
# ===========================================================================
# Pattern 1: Markdown table extraction
table_data = _extract_table_from_markdown(answer)
if table_data:
return _create_chart_from_table(table_data, chart_type, query, currency_symbol)
# Pattern 2: Numbered list with values (e.g., "1. Glovo Ltd - $1,153,500")
list_data = _extract_list_from_answer(answer)
if list_data:
return _create_chart_from_list(list_data, chart_type, query, currency_symbol)
# Pattern 3: Inline mentions (e.g., "Your top customer is Glovo Ltd with $1.15M")
inline_data = _extract_inline_data(answer, currency_symbol)
if inline_data:
return _create_chart_from_list(inline_data, chart_type, query, currency_symbol)
return None
def _extract_table_from_markdown(answer: str) -> Optional[List[Dict]]:
"""
Extract data from markdown tables in the answer.
Properly parses tables like:
| Customer | Revenue |
|----------|---------|
| Glovo Ltd | $1,153,500 |
"""
lines = answer.split('\n')
# Find table start (header row)
table_start = -1
for i, line in enumerate(lines):
# Table header has at least 2 pipe characters and no separator
if line.count('|') >= 2 and '---' not in line and '===' not in line:
# Check if next line is separator
if i + 1 < len(lines) and ('---' in lines[i + 1] or '===' in lines[i + 1]):
table_start = i
break
if table_start < 0:
return None
# Parse header
header_line = lines[table_start]
headers = []
for cell in header_line.split('|'):
cell = cell.strip()
# Clean markdown formatting
cell = re.sub(r'\*\*([^*]+)\*\*', r'\1', cell) # Remove bold
cell = re.sub(r'\*([^*]+)\*', r'\1', cell) # Remove italic
if cell and cell not in ['', '-', '---']:
headers.append(cell)
if len(headers) < 2:
return None
# Parse data rows (skip header and separator)
rows = []
for i in range(table_start + 2, len(lines)):
line = lines[i].strip()
# Stop at empty line or non-table line
if not line or '|' not in line:
break
# Parse cells
cells = []
for cell in line.split('|'):
cell = cell.strip()
# Clean markdown formatting
cell = re.sub(r'\*\*([^*]+)\*\*', r'\1', cell) # Remove bold
cell = re.sub(r'\*([^*]+)\*', r'\1', cell) # Remove italic
cell = cell.replace('\u202f', '') # Remove narrow non-breaking space
if cell and cell not in ['', '-', '---']:
cells.append(cell)
if len(cells) >= 2:
row_dict = {}
for j, header in enumerate(headers):
if j < len(cells):
row_dict[header] = cells[j]
if row_dict:
rows.append(row_dict)
# Validate we have actual data, not markdown artifacts
if rows:
# Check first row doesn't contain markdown syntax
first_row_values = list(rows[0].values())
for val in first_row_values:
if '|' in str(val) or '---' in str(val):
return None # Still has markdown artifacts
return rows if rows else None
def _extract_list_from_answer(answer: str) -> Optional[List[Dict]]:
"""Extract data from numbered or bulleted lists."""
# Pattern: "1. Name - $1,234,567" or "• Name: $1,234,567"
patterns = [
r'(?:\d+\.|•|-)\s*([^:\-–—]+?)[\s:\-–—]+[₹$€£]?([\d,]+(?:\.\d+)?)',
r'(?:\d+\.|•|-)\s*\*\*([^*]+)\*\*.*?[₹$€£]?([\d,]+(?:\.\d+)?)',
]
items = []
for pattern in patterns:
matches = re.findall(pattern, answer)
for match in matches:
name = match[0].strip()
value_str = match[1].replace(',', '')
try:
value = float(value_str)
if value > 0:
items.append({'name': name, 'value': value})
except ValueError:
continue
# Deduplicate by name
seen = set()
unique_items = []
for item in items:
if item['name'] not in seen:
seen.add(item['name'])
unique_items.append(item)
return unique_items[:10] if unique_items else None
def _extract_inline_data(answer: str, currency_symbol: str = "₹") -> Optional[List[Dict]]:
"""Extract entity-value pairs mentioned inline in the answer."""
# Pattern: "Customer_Name with $1,234,567" or "Customer_Name ($1,234,567)"
patterns = [
r'(\b[A-Z][a-zA-Z0-9_\s]+?)\s+(?:with|at|generated|earned)\s+[₹$€£]?([\d,]+(?:\.\d+)?)',
r'(\b[A-Z][a-zA-Z0-9_\s]+?)\s*\([₹$€£]?([\d,]+(?:\.\d+)?)\)',
r'\*\*([^*]+?)\*\*.*?[₹$€£]?([\d,]+(?:\.\d+)?)',
]
items = []
for pattern in patterns:
matches = re.findall(pattern, answer)
for match in matches:
name = match[0].strip()
# Skip common phrases
if name.lower() in ['your', 'the', 'total', 'revenue', 'amount', 'average']:
continue
value_str = match[1].replace(',', '')
try:
value = float(value_str)
if value > 0 and len(name) > 2:
items.append({'name': name, 'value': value})
except ValueError:
continue
return items[:10] if items else None
def _create_chart_from_table(
table_data: List[Dict],
chart_type: str,
query: str,
currency_symbol: str
) -> Dict:
"""Create chart specification from table data."""
if not table_data:
return None
# Find the name column (first text column)
name_col = None
value_col = None
first_row = table_data[0]
for key, val in first_row.items():
# Detect name column
if name_col is None and not _is_numeric(val):
name_col = key
# Detect value column (contains numbers)
elif _is_numeric(val):
value_col = key
if not name_col or not value_col:
return None
# Extract names and values
names = []
values = []
for row in table_data:
name = row.get(name_col, '')
value_str = row.get(value_col, '0')
try:
# Clean value string
value_clean = re.sub(r'[^\d.]', '', str(value_str))
value = float(value_clean) if value_clean else 0
names.append(name)
values.append(value)
except ValueError:
continue
return _build_plotly_chart(names, values, chart_type, query, currency_symbol)
def _create_chart_from_list(
list_data: List[Dict],
chart_type: str,
query: str,
currency_symbol: str
) -> Dict:
"""Create chart specification from list data."""
if not list_data:
return None
names = [item['name'] for item in list_data]
values = [item['value'] for item in list_data]
return _build_plotly_chart(names, values, chart_type, query, currency_symbol)
def _build_plotly_chart(
names: List[str],
values: List[float],
chart_type: str,
query: str,
currency_symbol: str
) -> Dict:
"""Build a Plotly chart specification."""
# Generate title from query
title = _generate_chart_title(query, names)
# Format values for display
def format_value(v):
if v >= 1_000_000:
return f"{currency_symbol}{v/1_000_000:.1f}M"
elif v >= 1_000:
return f"{currency_symbol}{v/1_000:.1f}K"
else:
return f"{currency_symbol}{v:,.0f}"
if chart_type == "pie":
return {
"data": [{
"type": "pie",
"labels": names,
"values": values,
"textinfo": "label+percent",
"hovertemplate": "%{label}<br>" + currency_symbol + "%{value:,.0f}<br>%{percent}<extra></extra>",
"marker": {
"colors": ['#FF6B35', '#4ECDC4', '#45B7D1', '#96CEB4', '#FFEAA7',
'#DDA0DD', '#98D8C8', '#F7DC6F', '#BB8FCE', '#85C1E9'][:len(names)]
}
}],
"layout": {
"title": {"text": title, "font": {"size": 16}},
"showlegend": True,
"height": 400
}
}
elif chart_type == "line":
return {
"data": [{
"type": "scatter",
"mode": "lines+markers",
"x": names,
"y": values,
"line": {"color": "#FF6B35", "width": 3},
"marker": {"size": 8}
}],
"layout": {
"title": {"text": title, "font": {"size": 16}},
"xaxis": {"title": ""},
"yaxis": {"title": "Amount", "tickprefix": currency_symbol},
"height": 400
}
}
else: # bar chart (default)
return {
"data": [{
"type": "bar",
"x": names,
"y": values,
"text": [format_value(v) for v in values],
"textposition": "auto",
"marker": {
"color": ['#FF6B35', '#4ECDC4', '#45B7D1', '#96CEB4', '#FFEAA7',
'#DDA0DD', '#98D8C8', '#F7DC6F', '#BB8FCE', '#85C1E9'][:len(names)]
}
}],
"layout": {
"title": {"text": title, "font": {"size": 16}},
"xaxis": {"title": "", "tickangle": -45},
"yaxis": {"title": "Amount", "tickprefix": currency_symbol},
"height": 400,
"margin": {"b": 100}
}
}
def _generate_chart_title(query: str, names: List[str]) -> str:
"""Generate a chart title from the query."""
query_lower = query.lower()
# Extract entity type from query
entity = "Items"
if "customer" in query_lower:
entity = "Customers"
elif "product" in query_lower:
entity = "Products"
elif "region" in query_lower:
entity = "Regions"
elif "category" in query_lower:
entity = "Categories"
# Detect top vs bottom
if "bottom" in query_lower or "low" in query_lower or "worst" in query_lower:
return f"📉 Bottom {len(names)} {entity} by Revenue"
elif "top" in query_lower or "best" in query_lower or "highest" in query_lower:
return f"🏆 Top {len(names)} {entity} by Revenue"
elif "vs" in query_lower or "compare" in query_lower:
return f"📊 {entity} Comparison"
else:
return f"📊 {entity} by Revenue"
def _is_numeric(value: str) -> bool:
"""Check if a value looks numeric."""
if not value:
return False
cleaned = re.sub(r'[₹$€£,%\s]', '', str(value))
try:
float(cleaned)
return True
except ValueError:
return False
|