File size: 23,303 Bytes
a9dc537 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 |
"""
LangChain-Compatible Tools for SPARKNET
All tools follow LangChain's tool interface for seamless integration
with LangGraph agents and workflows.
"""
from typing import Optional, List, Dict, Any
from pydantic import BaseModel, Field
from langchain_core.tools import StructuredTool, tool
from loguru import logger
import json
# PDF processing
try:
import PyPDF2
import fitz # pymupdf
PDF_AVAILABLE = True
except ImportError:
PDF_AVAILABLE = False
logger.warning("PDF libraries not available. Install PyPDF2 and pymupdf.")
# Document generation
try:
from reportlab.lib.pagesizes import letter
from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer
from reportlab.lib.styles import getSampleStyleSheet
REPORTLAB_AVAILABLE = True
except ImportError:
REPORTLAB_AVAILABLE = False
logger.warning("ReportLab not available. Install reportlab for PDF generation.")
# Web search and research
try:
from duckduckgo_search import DDGS
DDGS_AVAILABLE = True
except ImportError:
DDGS_AVAILABLE = False
logger.warning("DuckDuckGo search not available.")
try:
import wikipedia
WIKIPEDIA_AVAILABLE = True
except ImportError:
WIKIPEDIA_AVAILABLE = False
logger.warning("Wikipedia not available.")
try:
import arxiv
ARXIV_AVAILABLE = True
except ImportError:
ARXIV_AVAILABLE = False
logger.warning("Arxiv not available.")
# GPU monitoring
from ..utils.gpu_manager import get_gpu_manager
# ============================================================================
# Pydantic Input Schemas
# ============================================================================
class PDFExtractorInput(BaseModel):
"""Input schema for PDF extraction."""
file_path: str = Field(..., description="Path to the PDF file")
page_range: Optional[str] = Field(None, description="Page range (e.g., '1-5', 'all')")
extract_metadata: bool = Field(True, description="Extract PDF metadata")
class PatentParserInput(BaseModel):
"""Input schema for patent parsing."""
text: str = Field(..., description="Patent text to parse")
extract_claims: bool = Field(True, description="Extract patent claims")
extract_abstract: bool = Field(True, description="Extract abstract")
extract_description: bool = Field(True, description="Extract description")
class WebSearchInput(BaseModel):
"""Input schema for web search."""
query: str = Field(..., description="Search query")
max_results: int = Field(5, description="Maximum number of results")
region: str = Field("wt-wt", description="Search region (e.g., 'us-en', 'wt-wt')")
class WikipediaInput(BaseModel):
"""Input schema for Wikipedia lookup."""
query: str = Field(..., description="Wikipedia search query")
sentences: int = Field(3, description="Number of sentences to return")
class ArxivInput(BaseModel):
"""Input schema for Arxiv search."""
query: str = Field(..., description="Search query")
max_results: int = Field(5, description="Maximum number of results")
sort_by: str = Field("relevance", description="Sort by: relevance, lastUpdatedDate, submittedDate")
class DocumentGeneratorInput(BaseModel):
"""Input schema for document generation."""
output_path: str = Field(..., description="Output PDF file path")
title: str = Field(..., description="Document title")
content: str = Field(..., description="Document content (markdown or plain text)")
author: Optional[str] = Field(None, description="Document author")
class GPUMonitorInput(BaseModel):
"""Input schema for GPU monitoring."""
gpu_id: Optional[int] = Field(None, description="Specific GPU ID or None for all GPUs")
# ============================================================================
# PDF Tools
# ============================================================================
def pdf_extractor_func(file_path: str, page_range: Optional[str] = None,
extract_metadata: bool = True) -> str:
"""
Extract text and metadata from PDF files.
Supports both PyPDF2 and PyMuPDF (fitz) backends.
Args:
file_path: Path to PDF file
page_range: Page range like '1-5' or 'all' (default: all)
extract_metadata: Whether to extract metadata
Returns:
Extracted text and metadata as formatted string
"""
if not PDF_AVAILABLE:
return "Error: PDF libraries not installed. Run: pip install PyPDF2 pymupdf"
try:
# Open PDF with PyMuPDF (better text extraction)
doc = fitz.open(file_path)
# Parse page range
if page_range and page_range.lower() != 'all':
start, end = map(int, page_range.split('-'))
pages = range(start - 1, min(end, len(doc))) # 0-indexed
else:
pages = range(len(doc))
# Extract text
text_parts = []
for page_num in pages:
page = doc[page_num]
text_parts.append(f"--- Page {page_num + 1} ---\n{page.get_text()}")
extracted_text = "\n\n".join(text_parts)
# Extract metadata
result = f"PDF: {file_path}\n"
result += f"Total Pages: {len(doc)}\n"
result += f"Extracted Pages: {len(pages)}\n\n"
if extract_metadata:
metadata = doc.metadata
result += "Metadata:\n"
for key, value in metadata.items():
if value:
result += f" {key}: {value}\n"
result += "\n"
result += "=" * 80 + "\n"
result += "EXTRACTED TEXT:\n"
result += "=" * 80 + "\n"
result += extracted_text
doc.close()
logger.info(f"Extracted {len(pages)} pages from {file_path}")
return result
except Exception as e:
logger.error(f"PDF extraction failed: {e}")
return f"Error extracting PDF: {str(e)}"
def patent_parser_func(text: str, extract_claims: bool = True,
extract_abstract: bool = True, extract_description: bool = True) -> str:
"""
Parse patent document structure and extract key sections.
Uses heuristics to identify: abstract, claims, description, drawings.
Args:
text: Patent text (from PDF or plain text)
extract_claims: Extract patent claims
extract_abstract: Extract abstract
extract_description: Extract detailed description
Returns:
Structured patent information as JSON string
"""
try:
result = {
"abstract": "",
"claims": [],
"description": "",
"metadata": {}
}
lines = text.split('\n')
current_section = None
# Simple heuristic-based parser
for i, line in enumerate(lines):
line_lower = line.lower().strip()
# Detect sections
if 'abstract' in line_lower and len(line_lower) < 50:
current_section = 'abstract'
continue
elif 'claim' in line_lower and len(line_lower) < 50:
current_section = 'claims'
continue
elif 'description' in line_lower or 'detailed description' in line_lower:
if len(line_lower) < 100:
current_section = 'description'
continue
elif 'drawing' in line_lower or 'figure' in line_lower:
if len(line_lower) < 50:
current_section = 'drawings'
continue
# Extract content based on section
if current_section == 'abstract' and extract_abstract:
if line.strip():
result['abstract'] += line + "\n"
elif current_section == 'claims' and extract_claims:
if line.strip() and (line.strip()[0].isdigit() or 'wherein' in line_lower):
result['claims'].append(line.strip())
elif current_section == 'description' and extract_description:
if line.strip():
result['description'] += line + "\n"
# Extract patent number if present
for line in lines[:20]: # Check first 20 lines
if 'patent' in line.lower() and any(char.isdigit() for char in line):
result['metadata']['patent_number'] = line.strip()
break
# Format output
output = "PATENT ANALYSIS\n"
output += "=" * 80 + "\n\n"
if result['abstract']:
output += "ABSTRACT:\n"
output += result['abstract'].strip()[:500] # Limit length
output += "\n\n"
if result['claims']:
output += f"CLAIMS ({len(result['claims'])} found):\n"
for i, claim in enumerate(result['claims'][:10], 1): # First 10 claims
output += f"\n{i}. {claim}\n"
output += "\n"
if result['description']:
output += "DESCRIPTION (excerpt):\n"
output += result['description'].strip()[:1000] # First 1000 chars
output += "\n\n"
output += "=" * 80 + "\n"
output += f"JSON OUTPUT:\n{json.dumps(result, indent=2)}"
logger.info(f"Parsed patent: {len(result['claims'])} claims extracted")
return output
except Exception as e:
logger.error(f"Patent parsing failed: {e}")
return f"Error parsing patent: {str(e)}"
# ============================================================================
# Web Search & Research Tools
# ============================================================================
def web_search_func(query: str, max_results: int = 5, region: str = "wt-wt") -> str:
"""
Search the web using DuckDuckGo.
Returns top results with title, snippet, and URL.
Args:
query: Search query
max_results: Maximum number of results
region: Search region code
Returns:
Formatted search results
"""
if not DDGS_AVAILABLE:
return "Error: DuckDuckGo search not installed. Run: pip install duckduckgo-search"
try:
ddgs = DDGS()
results = list(ddgs.text(query, region=region, max_results=max_results))
if not results:
return f"No results found for: {query}"
output = f"WEB SEARCH RESULTS: {query}\n"
output += "=" * 80 + "\n\n"
for i, result in enumerate(results, 1):
output += f"{i}. {result.get('title', 'No title')}\n"
output += f" {result.get('body', 'No description')}\n"
output += f" URL: {result.get('href', 'No URL')}\n\n"
logger.info(f"Web search completed: {len(results)} results for '{query}'")
return output
except Exception as e:
logger.error(f"Web search failed: {e}")
return f"Error performing web search: {str(e)}"
def wikipedia_func(query: str, sentences: int = 3) -> str:
"""
Search Wikipedia and return summary.
Args:
query: Wikipedia search query
sentences: Number of sentences to return
Returns:
Wikipedia summary
"""
if not WIKIPEDIA_AVAILABLE:
return "Error: Wikipedia not installed. Run: pip install wikipedia"
try:
# Search for page
search_results = wikipedia.search(query)
if not search_results:
return f"No Wikipedia page found for: {query}"
# Get first result
page = wikipedia.page(search_results[0], auto_suggest=False)
# Get summary
summary = wikipedia.summary(search_results[0], sentences=sentences, auto_suggest=False)
output = f"WIKIPEDIA: {page.title}\n"
output += "=" * 80 + "\n\n"
output += summary + "\n\n"
output += f"URL: {page.url}\n"
output += f"Categories: {', '.join(page.categories[:5])}\n"
logger.info(f"Wikipedia lookup completed: {page.title}")
return output
except wikipedia.exceptions.DisambiguationError as e:
options = ', '.join(e.options[:5])
return f"Disambiguation needed for '{query}'. Options: {options}"
except wikipedia.exceptions.PageError:
return f"No Wikipedia page found for: {query}"
except Exception as e:
logger.error(f"Wikipedia lookup failed: {e}")
return f"Error: {str(e)}"
def arxiv_func(query: str, max_results: int = 5, sort_by: str = "relevance") -> str:
"""
Search Arxiv for academic papers.
Args:
query: Search query
max_results: Maximum number of results
sort_by: Sort by relevance, lastUpdatedDate, or submittedDate
Returns:
Formatted Arxiv results
"""
if not ARXIV_AVAILABLE:
return "Error: Arxiv not installed. Run: pip install arxiv"
try:
# Map sort_by to arxiv.SortCriterion
sort_map = {
"relevance": arxiv.SortCriterion.Relevance,
"lastUpdatedDate": arxiv.SortCriterion.LastUpdatedDate,
"submittedDate": arxiv.SortCriterion.SubmittedDate,
}
sort_criterion = sort_map.get(sort_by, arxiv.SortCriterion.Relevance)
# Search Arxiv
search = arxiv.Search(
query=query,
max_results=max_results,
sort_by=sort_criterion
)
results = list(search.results())
if not results:
return f"No Arxiv papers found for: {query}"
output = f"ARXIV SEARCH: {query}\n"
output += "=" * 80 + "\n\n"
for i, paper in enumerate(results, 1):
output += f"{i}. {paper.title}\n"
output += f" Authors: {', '.join(str(author) for author in paper.authors[:3])}\n"
output += f" Published: {paper.published.strftime('%Y-%m-%d')}\n"
output += f" Summary: {paper.summary[:200]}...\n"
output += f" PDF: {paper.pdf_url}\n"
output += f" Categories: {', '.join(paper.categories)}\n\n"
logger.info(f"Arxiv search completed: {len(results)} papers for '{query}'")
return output
except Exception as e:
logger.error(f"Arxiv search failed: {e}")
return f"Error searching Arxiv: {str(e)}"
# ============================================================================
# Document Generation
# ============================================================================
def document_generator_func(output_path: str, title: str, content: str,
author: Optional[str] = None) -> str:
"""
Generate PDF document from text content.
Supports basic formatting and styling.
Args:
output_path: Output PDF file path
title: Document title
content: Document content (plain text or simple markdown)
author: Optional author name
Returns:
Success message with file path
"""
if not REPORTLAB_AVAILABLE:
return "Error: ReportLab not installed. Run: pip install reportlab"
try:
# Create PDF
doc = SimpleDocTemplate(output_path, pagesize=letter)
styles = getSampleStyleSheet()
story = []
# Title
title_style = styles['Title']
story.append(Paragraph(title, title_style))
story.append(Spacer(1, 12))
# Author
if author:
author_style = styles['Normal']
story.append(Paragraph(f"By: {author}", author_style))
story.append(Spacer(1, 12))
# Content (split into paragraphs)
paragraphs = content.split('\n\n')
for para in paragraphs:
if para.strip():
# Simple markdown-like processing
if para.strip().startswith('#'):
# Heading
heading_text = para.strip().lstrip('#').strip()
story.append(Paragraph(heading_text, styles['Heading2']))
else:
# Regular paragraph
story.append(Paragraph(para.strip(), styles['Normal']))
story.append(Spacer(1, 6))
# Build PDF
doc.build(story)
logger.info(f"Generated PDF: {output_path}")
return f"Successfully generated PDF: {output_path}\nTitle: {title}\nPages: {len(paragraphs)}"
except Exception as e:
logger.error(f"PDF generation failed: {e}")
return f"Error generating PDF: {str(e)}"
# ============================================================================
# GPU Monitoring (converted from existing tool)
# ============================================================================
def gpu_monitor_func(gpu_id: Optional[int] = None) -> str:
"""
Monitor GPU status, memory usage, and utilization.
Args:
gpu_id: Specific GPU ID or None for all GPUs
Returns:
Formatted GPU status information
"""
try:
gpu_manager = get_gpu_manager()
if gpu_id is not None:
# Monitor specific GPU
info = gpu_manager.get_gpu_info(gpu_id)
if "error" in info:
return f"Error: {info['error']}"
output = f"GPU {info['gpu_id']}: {info['name']}\n"
output += f"Memory: {info['memory_used'] / 1024**3:.2f} GB / {info['memory_total'] / 1024**3:.2f} GB "
output += f"({info['memory_percent']:.1f}% used)\n"
output += f"Free Memory: {info['memory_free'] / 1024**3:.2f} GB\n"
output += f"GPU Utilization: {info['gpu_utilization']}%\n"
output += f"Temperature: {info['temperature']}°C\n"
return output
else:
# Monitor all GPUs
return gpu_manager.monitor()
except Exception as e:
logger.error(f"GPU monitoring error: {e}")
return f"Error monitoring GPU: {str(e)}"
# ============================================================================
# Create LangChain Tools
# ============================================================================
# Use StructuredTool for tools with Pydantic input schemas
pdf_extractor_tool = StructuredTool.from_function(
func=pdf_extractor_func,
name="pdf_extractor",
description=(
"Extract text and metadata from PDF files. "
"Useful for analyzing patent documents, research papers, and legal documents. "
"Supports page range selection and metadata extraction."
),
args_schema=PDFExtractorInput,
return_direct=False,
)
patent_parser_tool = StructuredTool.from_function(
func=patent_parser_func,
name="patent_parser",
description=(
"Parse patent document structure and extract key sections: abstract, claims, description. "
"Useful for analyzing patent documents and identifying key innovations."
),
args_schema=PatentParserInput,
return_direct=False,
)
web_search_tool = StructuredTool.from_function(
func=web_search_func,
name="web_search",
description=(
"Search the web using DuckDuckGo. Returns top results with titles, snippets, and URLs. "
"Useful for market research, competitor analysis, and finding relevant information."
),
args_schema=WebSearchInput,
return_direct=False,
)
wikipedia_tool = StructuredTool.from_function(
func=wikipedia_func,
name="wikipedia",
description=(
"Search Wikipedia and get article summaries. "
"Useful for background information on technologies, companies, and concepts."
),
args_schema=WikipediaInput,
return_direct=False,
)
arxiv_tool = StructuredTool.from_function(
func=arxiv_func,
name="arxiv_search",
description=(
"Search Arxiv for academic papers and preprints. "
"Useful for finding relevant research, state-of-the-art methods, and technical background."
),
args_schema=ArxivInput,
return_direct=False,
)
document_generator_tool = StructuredTool.from_function(
func=document_generator_func,
name="document_generator",
description=(
"Generate PDF documents from text content. "
"Useful for creating reports, briefs, and documentation."
),
args_schema=DocumentGeneratorInput,
return_direct=False,
)
gpu_monitor_tool = StructuredTool.from_function(
func=gpu_monitor_func,
name="gpu_monitor",
description=(
"Monitor GPU status including memory usage, utilization, and temperature. "
"Useful for checking GPU availability before running models."
),
args_schema=GPUMonitorInput,
return_direct=False,
)
# ============================================================================
# Tool Registry for VISTA Scenarios
# ============================================================================
class VISTAToolRegistry:
"""
Registry of tools organized by VISTA scenario.
Enables scenario-specific tool selection for optimal performance.
"""
SCENARIO_TOOLS = {
"patent_wakeup": [
pdf_extractor_tool,
patent_parser_tool,
web_search_tool,
wikipedia_tool,
arxiv_tool,
document_generator_tool,
],
"agreement_safety": [
pdf_extractor_tool,
web_search_tool,
document_generator_tool,
],
"partner_matching": [
web_search_tool,
wikipedia_tool,
arxiv_tool,
],
"general": [
pdf_extractor_tool,
patent_parser_tool,
web_search_tool,
wikipedia_tool,
arxiv_tool,
document_generator_tool,
gpu_monitor_tool,
],
}
@classmethod
def get_tools(cls, scenario: str = "general") -> List[StructuredTool]:
"""
Get tools for a specific VISTA scenario.
Args:
scenario: VISTA scenario type
Returns:
List of LangChain tools
"""
tools = cls.SCENARIO_TOOLS.get(scenario, cls.SCENARIO_TOOLS["general"])
logger.info(f"Retrieved {len(tools)} tools for scenario: {scenario}")
return tools
@classmethod
def get_all_tools(cls) -> List[StructuredTool]:
"""Get all available tools."""
return cls.SCENARIO_TOOLS["general"]
@classmethod
def list_scenarios(cls) -> List[str]:
"""List available scenarios."""
return list(cls.SCENARIO_TOOLS.keys())
# ============================================================================
# Convenience Functions
# ============================================================================
def get_vista_tools(scenario: str = "general") -> List[StructuredTool]:
"""
Get LangChain tools for a VISTA scenario.
Args:
scenario: Scenario name (patent_wakeup, agreement_safety, partner_matching, general)
Returns:
List of LangChain StructuredTool instances
"""
return VISTAToolRegistry.get_tools(scenario)
def get_all_tools() -> List[StructuredTool]:
"""Get all available LangChain tools."""
return VISTAToolRegistry.get_all_tools()
# Export all tools
__all__ = [
"pdf_extractor_tool",
"patent_parser_tool",
"web_search_tool",
"wikipedia_tool",
"arxiv_tool",
"document_generator_tool",
"gpu_monitor_tool",
"VISTAToolRegistry",
"get_vista_tools",
"get_all_tools",
]
|