local-deep-research 0.2.2__py3-none-any.whl → 0.3.0__py3-none-any.whl

local_deep_research/advanced_search_system/strategies/source_based_strategy.py

@@ -0,0 +1,402 @@
+import concurrent.futures
+import logging
+from typing import Dict
+
+from ...citation_handler import CitationHandler
+from ...config.llm_config import get_llm
+from ...config.search_config import get_search
+from ...utilities.db_utils import get_db_setting
+from ..filters.cross_engine_filter import CrossEngineFilter
+from ..findings.repository import FindingsRepository
+from ..questions.standard_question import StandardQuestionGenerator
+from .base_strategy import BaseSearchStrategy
+
+logger = logging.getLogger(__name__)
+
+
+class SourceBasedSearchStrategy(BaseSearchStrategy):
+    """
+    Source-based search strategy that generates questions based on search results and
+    defers content analysis until final synthesis.
+    """
+
+    def __init__(
+        self,
+        search=None,
+        model=None,
+        citation_handler=None,
+        include_text_content: bool = True,
+        use_cross_engine_filter: bool = True,
+        filter_reorder: bool = True,
+        filter_reindex: bool = True,
+        cross_engine_max_results: int = None,
+        all_links_of_system=None,
+    ):
+        """Initialize with optional dependency injection for testing."""
+        # Pass the links list to the parent class
+        super().__init__(all_links_of_system=all_links_of_system)
+        self.search = search or get_search()
+        self.model = model or get_llm()
+        self.progress_callback = None
+
+        self.questions_by_iteration = {}
+        self.include_text_content = include_text_content
+        self.use_cross_engine_filter = use_cross_engine_filter
+        self.filter_reorder = filter_reorder
+        self.filter_reindex = filter_reindex
+
+        # Get cross_engine_max_results from database if not provided
+        if cross_engine_max_results is None:
+            cross_engine_max_results = get_db_setting(
+                "search.cross_engine_max_results", 100
+            )
+
+        # Initialize the cross-engine filter
+        self.cross_engine_filter = CrossEngineFilter(
+            model=self.model,
+            max_results=cross_engine_max_results,
+            default_reorder=filter_reorder,
+            default_reindex=filter_reindex,
+        )
+
+        # Set include_full_content on the search engine if it supports it
+        if hasattr(self.search, "include_full_content"):
+            self.search.include_full_content = include_text_content
+
+        # Use provided citation_handler or create one
+        self.citation_handler = citation_handler or CitationHandler(self.model)
+
+        # Initialize components
+        self.question_generator = StandardQuestionGenerator(self.model)
+        self.findings_repository = FindingsRepository(self.model)
+
+    def _format_search_results_as_context(self, search_results):
+        """Format search results into context for question generation."""
+        context_snippets = []
+
+        for i, result in enumerate(
+            search_results[:10]
+        ):  # Limit to prevent context overflow
+            title = result.get("title", "Untitled")
+            snippet = result.get("snippet", "")
+            url = result.get("link", "")
+
+            if snippet:
+                context_snippets.append(
+                    f"Source {i + 1}: {title}\nURL: {url}\nSnippet: {snippet}"
+                )
+
+        return "\n\n".join(context_snippets)
+
+    def analyze_topic(self, query: str) -> Dict:
+        """
+        Analyze a topic using source-based search strategy.
+        """
+        logger.info(f"Starting source-based research on topic: {query}")
+        accumulated_search_results_across_all_iterations = (
+            []
+        )  # tracking links across iterations but not global
+        findings = []
+        total_citation_count_before_this_search = len(self.all_links_of_system)
+
+        self._update_progress(
+            "Initializing source-based research",
+            5,
+            {
+                "phase": "init",
+                "strategy": "source-based",
+                "include_text_content": self.include_text_content,
+            },
+        )
+
+        # Check search engine
+        if not self._validate_search_engine():
+            return {
+                "findings": [],
+                "iterations": 0,
+                "questions_by_iteration": {},
+                "formatted_findings": "Error: Unable to conduct research without a search engine.",
+                "current_knowledge": "",
+                "error": "No search engine available",
+            }
+
+        # Determine number of iterations to run
+        iterations_to_run = get_db_setting("search.iterations")
+        logger.debug("Selected amount of iterations: " + str(iterations_to_run))
+        iterations_to_run = int(iterations_to_run)
+        try:
+            filtered_search_results = []
+            total_citation_count_before_this_search = len(self.all_links_of_system)
+            # Run each iteration
+            for iteration in range(1, iterations_to_run + 1):
+                iteration_progress_base = 5 + (iteration - 1) * (70 / iterations_to_run)
+
+                self._update_progress(
+                    f"Starting iteration {iteration}/{iterations_to_run}",
+                    iteration_progress_base,
+                    {"phase": f"iteration_{iteration}", "iteration": iteration},
+                )
+
+                # Step 1: Generate or use questions
+                self._update_progress(
+                    f"Generating search questions for iteration {iteration}",
+                    iteration_progress_base + 5,
+                    {"phase": "question_generation", "iteration": iteration},
+                )
+
+                # For first iteration, use initial query
+                if iteration == 1:
+                    # Generate questions for first iteration
+                    context = f"""Iteration: {iteration} of {iterations_to_run}"""
+                    questions = self.question_generator.generate_questions(
+                        current_knowledge=context,
+                        query=query,
+                        questions_per_iteration=int(
+                            get_db_setting("search.questions_per_iteration")
+                        ),
+                        questions_by_iteration=self.questions_by_iteration,
+                    )
+
+                    # Always include the original query for the first iteration
+                    if query not in questions:
+                        all_questions = [query] + questions
+                    else:
+                        all_questions = questions
+
+                    self.questions_by_iteration[iteration] = all_questions
+                    logger.info(
+                        f"Using questions for iteration {iteration}: {all_questions}"
+                    )
+                else:
+                    # For subsequent iterations, generate questions based on previous search results
+                    source_context = self._format_search_results_as_context(
+                        filtered_search_results
+                    )
+                    if iteration != 1:
+                        context = f"""Previous search results:\n{source_context}\n\nIteration: {iteration} of {iterations_to_run}"""
+                    elif iterations_to_run == 1:
+                        context = ""
+                    else:
+                        context = f"""Iteration: {iteration} of {iterations_to_run}"""
+                    # Use standard question generator with search results as context
+                    questions = self.question_generator.generate_questions(
+                        current_knowledge=context,
+                        query=query,
+                        questions_per_iteration=int(
+                            get_db_setting("search.questions_per_iteration")
+                        ),
+                        questions_by_iteration=self.questions_by_iteration,
+                    )
+
+                    # Use only the new questions for this iteration's searches
+                    all_questions = questions
+
+                    # Store in questions_by_iteration
+                    self.questions_by_iteration[iteration] = questions
+                    logger.info(
+                        f"Generated questions for iteration {iteration}: {questions}"
+                    )
+
+                # Step 2: Run all searches in parallel for this iteration
+                self._update_progress(
+                    f"Running parallel searches for iteration {iteration}",
+                    iteration_progress_base + 10,
+                    {"phase": "parallel_search", "iteration": iteration},
+                )
+
+                # Function for thread pool
+                def search_question(q):
+                    try:
+                        result = self.search.run(q)
+                        return {"question": q, "results": result or []}
+                    except Exception as e:
+                        logger.error(f"Error searching for '{q}': {str(e)}")
+                        return {"question": q, "results": [], "error": str(e)}
+
+                # Run searches in parallel
+                with concurrent.futures.ThreadPoolExecutor(
+                    max_workers=len(all_questions)
+                ) as executor:
+                    futures = [
+                        executor.submit(search_question, q) for q in all_questions
+                    ]
+                    iteration_search_dict = {}
+                    iteration_search_results = []
+
+                    # Process results as they complete
+                    for i, future in enumerate(
+                        concurrent.futures.as_completed(futures)
+                    ):
+                        result_dict = future.result()
+                        question = result_dict["question"]
+                        search_results = result_dict["results"]
+                        iteration_search_dict[question] = search_results
+
+                        self._update_progress(
+                            f"Completed search {i + 1} of {len(all_questions)}: {question[:30]}...",
+                            iteration_progress_base
+                            + 10
+                            + ((i + 1) / len(all_questions) * 30),
+                            {
+                                "phase": "search_complete",
+                                "iteration": iteration,
+                                "result_count": len(search_results),
+                                "question": question,
+                            },
+                        )
+
+                        iteration_search_results.extend(search_results)
+
+                if False and self.use_cross_engine_filter:
+                    self._update_progress(
+                        f"Filtering search results for iteration {iteration}",
+                        iteration_progress_base + 45,
+                        {"phase": "cross_engine_filtering", "iteration": iteration},
+                    )
+
+                    existing_link_count = len(self.all_links_of_system)
+                    logger.info(f"Existing link count: {existing_link_count}")
+                    filtered_search_results = self.cross_engine_filter.filter_results(
+                        iteration_search_results,
+                        query,
+                        reorder=True,
+                        reindex=True,
+                        start_index=existing_link_count,  # Start indexing after existing links
+                    )
+
+                    self._update_progress(
+                        f"Filtered from {len(iteration_search_results)} to {len(filtered_search_results)} results",
+                        iteration_progress_base + 50,
+                        {
+                            "phase": "filtering_complete",
+                            "iteration": iteration,
+                            "links_count": len(self.all_links_of_system),
+                        },
+                    )
+                else:
+                    # Use the search results as they are
+                    filtered_search_results = iteration_search_results
+
+                    # Use filtered results
+                accumulated_search_results_across_all_iterations.extend(
+                    filtered_search_results
+                )
+
+                # Create a lightweight finding for this iteration's search metadata (no text content)
+                finding = {
+                    "phase": f"Iteration {iteration}",
+                    "content": f"Searched with {len(all_questions)} questions, found {len(filtered_search_results)} results.",
+                    "question": query,
+                    "documents": [],
+                }
+                findings.append(finding)
+
+                # Mark iteration as complete
+                iteration_progress = 5 + iteration * (70 / iterations_to_run)
+                self._update_progress(
+                    f"Completed iteration {iteration}/{iterations_to_run}",
+                    iteration_progress,
+                    {"phase": "iteration_complete", "iteration": iteration},
+                )
+
+            # Do we need this filter?
+            if self.use_cross_engine_filter:
+                # Final filtering of all accumulated search results
+                self._update_progress(
+                    "Performing final filtering of all results",
+                    80,
+                    {"phase": "final_filtering"},
+                )
+                final_filtered_results = self.cross_engine_filter.filter_results(
+                    accumulated_search_results_across_all_iterations,
+                    query,
+                    reorder=True,  # Always reorder in final filtering
+                    reindex=True,  # Always reindex in final filtering
+                    max_results=int(get_db_setting("search.final_max_results") or 100),
+                    start_index=len(self.all_links_of_system),
+                )
+                self._update_progress(
+                    f"Filtered from {len(accumulated_search_results_across_all_iterations)} to {len(final_filtered_results)} results",
+                    iteration_progress_base + 85,
+                    {
+                        "phase": "filtering_complete",
+                        "iteration": iteration,
+                        "links_count": len(self.all_links_of_system),
+                    },
+                )
+            else:
+                final_filtered_results = filtered_search_results
+                # links = extract_links_from_search_results()
+            self.all_links_of_system.extend(final_filtered_results)
+
+            # Final synthesis after all iterations
+            self._update_progress(
+                "Generating final synthesis", 90, {"phase": "synthesis"}
+            )
+
+            # Final synthesis
+            final_citation_result = self.citation_handler.analyze_followup(
+                query,
+                final_filtered_results,
+                previous_knowledge="",  # Empty string as we don't need previous knowledge here
+                nr_of_links=total_citation_count_before_this_search,
+            )
+
+            # Add null check for final_citation_result
+            if final_citation_result:
+                synthesized_content = final_citation_result["content"]
+                documents = final_citation_result.get("documents", [])
+            else:
+                synthesized_content = "No relevant results found in final synthesis."
+                documents = []
+
+            # Add a final synthesis finding
+            final_finding = {
+                "phase": "Final synthesis",
+                "content": synthesized_content,
+                "question": query,
+                "search_results": self.all_links_of_system,
+                "documents": documents,
+            }
+            findings.append(final_finding)
+
+            # Add documents to repository
+            self.findings_repository.add_documents(documents)
+
+            # Transfer questions to repository
+            self.findings_repository.set_questions_by_iteration(
+                self.questions_by_iteration
+            )
+
+            # Format findings
+            formatted_findings = self.findings_repository.format_findings_to_text(
+                findings, synthesized_content
+            )
+
+        except Exception as e:
+            import traceback
+
+            error_msg = f"Error in research process: {str(e)}"
+            logger.error(error_msg)
+            logger.error(traceback.format_exc())
+            synthesized_content = f"Error: {str(e)}"
+            formatted_findings = f"Error: {str(e)}"
+            finding = {
+                "phase": "Error",
+                "content": synthesized_content,
+                "question": query,
+                "search_results": [],
+                "documents": [],
+            }
+            findings.append(finding)
+
+        self._update_progress("Research complete", 100, {"phase": "complete"})
+
+        return {
+            "findings": findings,
+            "iterations": iterations_to_run,
+            "questions_by_iteration": self.questions_by_iteration,
+            "formatted_findings": formatted_findings,
+            "current_knowledge": synthesized_content,
+            "all_links_of_system": self.all_links_of_system,
+        }

local_deep_research/advanced_search_system/strategies/standard_strategy.py

@@ -3,7 +3,6 @@ import logging
 from typing import Dict
 
 from ...citation_handler import CitationHandler
-from ...config.config_files import settings
 from ...config.llm_config import get_llm
 from ...config.search_config import get_search
 from ...utilities.db_utils import get_db_setting
@@ -20,11 +19,17 @@ logger = logging.getLogger(__name__)
 class StandardSearchStrategy(BaseSearchStrategy):
     """Standard iterative search strategy that generates follow-up questions."""
 
-    def __init__(self, search=None, model=None, citation_handler=None):
+    def __init__(
+        self, search=None, model=None, citation_handler=None, all_links_of_system=None
+    ):
         """Initialize with optional dependency injection for testing."""
+        super().__init__(all_links_of_system=all_links_of_system)
         self.search = search or get_search()
         self.model = model or get_llm()
+
+        # Get iterations setting
         self.max_iterations = int(get_db_setting("search.iterations"))
+
         self.questions_per_iteration = int(
             get_db_setting("search.questions_per_iteration")
         )
@@ -43,7 +48,6 @@ class StandardSearchStrategy(BaseSearchStrategy):
 
         # Initialize other attributes
         self.progress_callback = None
-        self.all_links_of_system = list()
 
     def _update_progress(
         self, message: str, progress_percent: int = None, metadata: dict = None
@@ -117,7 +121,7 @@ Iteration: {iteration + 1} of {total_iterations}"""
             question_count = len(questions)
             knowledge_accumulation = get_db_setting(
                 "general.knowledge_accumulation",
-                settings.general.knowledge_accumulation,
+                "ITERATION",
             )
             for q_idx, question in enumerate(questions):
                 question_progress_base = iteration_progress_base + (

local_deep_research/api/research_functions.py

@@ -4,11 +4,8 @@ Provides programmatic access to search and research capabilities.
 """
 
 import logging
-import os
 from typing import Any, Callable, Dict, Optional
 
-import toml
-
 from ..config.llm_config import get_llm
 from ..config.search_config import get_search
 from ..report_generator import IntegratedReportGenerator
@@ -279,46 +276,3 @@ def analyze_documents(
         logger.info(f"Analysis saved to {output_file}")
 
     return analysis_result
-
-
-def get_available_search_engines() -> Dict[str, str]:
-    """
-    Get a dictionary of available search engines.
-
-    Returns:
-        Dictionary mapping engine names to descriptions
-    """
-
-    from ..web_search_engines.search_engine_factory import get_available_engines
-
-    engines = get_available_engines()
-
-    # Add some descriptions for common engines
-    descriptions = {
-        "auto": "Automatic selection based on query type",
-        "wikipedia": "Wikipedia articles and general knowledge",
-        "arxiv": "Scientific papers and research",
-        "pubmed": "Medical and biomedical literature",
-        "semantic_scholar": "Academic papers across all fields",
-        "github": "Code repositories and technical documentation",
-        "local_all": "All local document collections",
-    }
-
-    return {engine: descriptions.get(engine, "Search engine") for engine in engines}
-
-
-def get_available_collections() -> Dict[str, Dict[str, Any]]:
-    """
-    Get a dictionary of available local document collections.
-
-    Returns:
-        Dictionary mapping collection names to their configuration
-    """
-
-    from ..config.config_files import LOCAL_COLLECTIONS_FILE
-
-    if os.path.exists(LOCAL_COLLECTIONS_FILE):
-        collections = toml.load(LOCAL_COLLECTIONS_FILE)
-        return collections
-
-    return {}

local_deep_research/citation_handler.py

@@ -4,7 +4,6 @@ from typing import Any, Dict, List, Union
 
 from langchain_core.documents import Document
 
-from .config.config_files import settings
 from .utilities.db_utils import get_db_setting
 
 
@@ -82,21 +81,19 @@ Provide a detailed analysis with citations. Do not create the bibliography, it w
         formatted_sources = self._format_sources(documents)
         # Add fact-checking step
         fact_check_prompt = f"""Analyze these sources for factual consistency:
-        1. Cross-reference major claims between sources
-        2. Identify and flag any contradictions
-        3. Verify basic facts (dates, company names, ownership)
-        4. Note when sources disagree
+1. Cross-reference major claims between sources
+2. Identify and flag any contradictions
+3. Verify basic facts (dates, company names, ownership)
+4. Note when sources disagree
 
-        Previous Knowledge:
-        {previous_knowledge}
+Previous Knowledge:
+{previous_knowledge}
 
-        New Sources:
-        {formatted_sources}
+New Sources:
+{formatted_sources}
 
         Return any inconsistencies or conflicts found."""
-        if get_db_setting(
-            "general.enable_fact_checking", settings.general.enable_fact_checking
-        ):
+        if get_db_setting("general.enable_fact_checking", True):
             fact_check_response = self.llm.invoke(fact_check_prompt).content
 
         else:
@@ -104,16 +101,15 @@ Provide a detailed analysis with citations. Do not create the bibliography, it w
 
         prompt = f"""Using the previous knowledge and new sources, answer the question. Include citations using numbers in square brackets [1], [2], etc. When citing, use the source number provided at the start of each source. Reflect information from sources critically.
 
-            Previous Knowledge:
-            {previous_knowledge}
+Previous Knowledge:
+{previous_knowledge}
 
-            Question: {question}
+Question: {question}
 
-            New Sources:
-            {formatted_sources}
-            Reflect information from sources critically based on: {fact_check_response}. Never invent sources.
-            Provide a detailed answer with citations.  Example format: "According to [1], ..."
-            """
+New Sources:
+{formatted_sources}
+Reflect information from sources critically based on: {fact_check_response}. Never invent sources.
+Provide a detailed answer with citations.  Example format: "According to [1], ..." """
 
         response = self.llm.invoke(prompt)