biomapper 0.3.1__tar.gz → 0.3.2__tar.gz

{biomapper-0.3.1 → biomapper-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: biomapper
-Version: 0.3.1
+Version: 0.3.2
 Summary: A unified Python toolkit for biological data harmonization and ontology mapping
 Home-page: https://github.com/arpanauts/biomapper
 License: MIT
@@ -20,9 +20,14 @@ Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
 Provides-Extra: api
 Provides-Extra: full
+Requires-Dist: cloudpickle (>=3.0.0,<4.0.0)
+Requires-Dist: dspy-ai (>=2.1.8,<3.0.0)
+Requires-Dist: langfuse (>=2.57.1,<3.0.0)
 Requires-Dist: libChEBIpy (==1.0.10)
 Requires-Dist: matplotlib (>=3.8.0,<4.0.0)
+Requires-Dist: openai (>=1.14.0,<2.0.0)
 Requires-Dist: pandas (>=2.0.0,<3.0.0)
+Requires-Dist: python-dotenv (>=1.0.1,<2.0.0)
 Requires-Dist: pyyaml (>=6.0.1,<7.0.0)
 Requires-Dist: requests (>=2.25.1,<3.0.0)
 Requires-Dist: seaborn (>=0.13.0,<0.14.0)
@@ -42,21 +47,22 @@ A unified Python toolkit for biological data harmonization and ontology mapping.
 
 ### Core Functionality
 - **ID Standardization**: Unified interface for standardizing biological identifiers
-- **Ontology Mapping**: Comprehensive ontology mapping using major biological databases
+- **Ontology Mapping**: Comprehensive ontology mapping using major biological databases and AI-powered techniques
 - **Data Validation**: Robust validation of input data and mappings
 - **Extensible Architecture**: Easy integration of new data sources and mapping services
 
 ### Supported Systems
 
 #### ID Standardization Tools
-- BridgeDb
-- RefMet
-- RaMP-DB
+- RaMP-DB: Integration with the Rapid Mapping Database for metabolites and pathways
 
-#### Ontology Mapping Services
-- UMLS Metathesaurus
-- Ontology Lookup Service (OLS)
-- BioPortal
+#### Mapping Services
+- ChEBI: Chemical Entities of Biological Interest database integration
+- UniChem: Cross-referencing of chemical structure identifiers
+- UniProt: Protein-focused mapping capabilities
+- RefMet: Reference list of metabolite names and identifiers
+- RAG-Based Mapping: AI-powered mapping using Retrieval Augmented Generation
+- Multi-Provider RAG: Combining multiple data sources for improved mapping accuracy
 
 ## Installation
 
@@ -112,20 +118,18 @@ poetry install
 ## Quick Start
 
 ```python
-from biomapper import AnalyteMetadata
-from biomapper.standardization import BridgeDBHandler, RaMPClient
+from biomapper.mapping import UniProtFocusedMapper, MetaboliteNameMapper
+from biomapper.standardization import RaMPClient
 
-# Example 1: Using BridgeDB
-# Initialize metadata handler
-metadata = AnalyteMetadata()
+# Example 1: Using UniProt-focused mapping
+uniprot_mapper = UniProtFocusedMapper()
+protein_mapping = uniprot_mapper.map_identifier("P12345")
 
-# Create standardization handler
-bridge_handler = BridgeDBHandler()
+# Example 2: Using Metabolite Name Mapping
+metabolite_mapper = MetaboliteNameMapper()
+metabolite_mapping = metabolite_mapper.map_name("glucose")
 
-# Process identifiers
-results = bridge_handler.standardize(["P12345", "Q67890"])
-
-# Example 2: Using RaMP-DB
+# Example 3: Using RaMP-DB
 # Initialize the RaMP client
 ramp_client = RaMPClient()
 
@@ -135,6 +139,12 @@ versions = ramp_client.get_source_versions()
 # Get pathways for metabolites
 # Example: Get pathways for Creatine (HMDB0000064)
 pathways = ramp_client.get_pathways_from_analytes(["hmdb:HMDB0000064"])
+
+# Example 4: Using RAG-based mapping
+from biomapper.mapping import RagMapper
+
+rag_mapper = RagMapper()
+rag_results = rag_mapper.map_name("alpha-D-glucose")
 ```
 
 ## Development
@@ -214,17 +224,18 @@ For support, please open an issue in the GitHub issue tracker.
 
 ## Roadmap
 
-- [ ] Initial release with core functionality
-- [ ] Add support for additional ontology services
-- [ ] Implement caching layer
+- [x] Initial release with core functionality
+- [x] Implement RAG-based mapping capabilities
+- [x] Add support for major chemical/biological databases (ChEBI, UniChem, UniProt)
+- [ ] Add caching layer for improved performance
+- [ ] Expand RAG capabilities with more specialized models
 - [ ] Add batch processing capabilities
 - [ ] Develop REST API interface
 
 ## Acknowledgments
 
-- [BridgeDb](https://www.bridgedb.org/)
-- [RefMet](https://refmet.metabolomicsworkbench.org/)
 - [RaMP-DB](http://rampdb.org/)
-- [UMLS](https://www.nlm.nih.gov/research/umls/index.html)
-- [OLS](https://www.ebi.ac.uk/ols/index)
-- [BioPortal](https://bioportal.bioontology.org/)
+- [ChEBI](https://www.ebi.ac.uk/chebi/)
+- [UniChem](https://www.ebi.ac.uk/unichem/)
+- [UniProt](https://www.uniprot.org/)
+- [RefMet](https://refmet.metabolomicsworkbench.org/)

{biomapper-0.3.1 → biomapper-0.3.2}/README.md

@@ -6,21 +6,22 @@ A unified Python toolkit for biological data harmonization and ontology mapping.
 
 ### Core Functionality
 - **ID Standardization**: Unified interface for standardizing biological identifiers
-- **Ontology Mapping**: Comprehensive ontology mapping using major biological databases
+- **Ontology Mapping**: Comprehensive ontology mapping using major biological databases and AI-powered techniques
 - **Data Validation**: Robust validation of input data and mappings
 - **Extensible Architecture**: Easy integration of new data sources and mapping services
 
 ### Supported Systems
 
 #### ID Standardization Tools
-- BridgeDb
-- RefMet
-- RaMP-DB
+- RaMP-DB: Integration with the Rapid Mapping Database for metabolites and pathways
 
-#### Ontology Mapping Services
-- UMLS Metathesaurus
-- Ontology Lookup Service (OLS)
-- BioPortal
+#### Mapping Services
+- ChEBI: Chemical Entities of Biological Interest database integration
+- UniChem: Cross-referencing of chemical structure identifiers
+- UniProt: Protein-focused mapping capabilities
+- RefMet: Reference list of metabolite names and identifiers
+- RAG-Based Mapping: AI-powered mapping using Retrieval Augmented Generation
+- Multi-Provider RAG: Combining multiple data sources for improved mapping accuracy
 
 ## Installation
 
@@ -76,20 +77,18 @@ poetry install
 ## Quick Start
 
 ```python
-from biomapper import AnalyteMetadata
-from biomapper.standardization import BridgeDBHandler, RaMPClient
+from biomapper.mapping import UniProtFocusedMapper, MetaboliteNameMapper
+from biomapper.standardization import RaMPClient
 
-# Example 1: Using BridgeDB
-# Initialize metadata handler
-metadata = AnalyteMetadata()
+# Example 1: Using UniProt-focused mapping
+uniprot_mapper = UniProtFocusedMapper()
+protein_mapping = uniprot_mapper.map_identifier("P12345")
 
-# Create standardization handler
-bridge_handler = BridgeDBHandler()
+# Example 2: Using Metabolite Name Mapping
+metabolite_mapper = MetaboliteNameMapper()
+metabolite_mapping = metabolite_mapper.map_name("glucose")
 
-# Process identifiers
-results = bridge_handler.standardize(["P12345", "Q67890"])
-
-# Example 2: Using RaMP-DB
+# Example 3: Using RaMP-DB
 # Initialize the RaMP client
 ramp_client = RaMPClient()
 
@@ -99,6 +98,12 @@ versions = ramp_client.get_source_versions()
 # Get pathways for metabolites
 # Example: Get pathways for Creatine (HMDB0000064)
 pathways = ramp_client.get_pathways_from_analytes(["hmdb:HMDB0000064"])
+
+# Example 4: Using RAG-based mapping
+from biomapper.mapping import RagMapper
+
+rag_mapper = RagMapper()
+rag_results = rag_mapper.map_name("alpha-D-glucose")
 ```
 
 ## Development
@@ -178,17 +183,18 @@ For support, please open an issue in the GitHub issue tracker.
 
 ## Roadmap
 
-- [ ] Initial release with core functionality
-- [ ] Add support for additional ontology services
-- [ ] Implement caching layer
+- [x] Initial release with core functionality
+- [x] Implement RAG-based mapping capabilities
+- [x] Add support for major chemical/biological databases (ChEBI, UniChem, UniProt)
+- [ ] Add caching layer for improved performance
+- [ ] Expand RAG capabilities with more specialized models
 - [ ] Add batch processing capabilities
 - [ ] Develop REST API interface
 
 ## Acknowledgments
 
-- [BridgeDb](https://www.bridgedb.org/)
-- [RefMet](https://refmet.metabolomicsworkbench.org/)
 - [RaMP-DB](http://rampdb.org/)
-- [UMLS](https://www.nlm.nih.gov/research/umls/index.html)
-- [OLS](https://www.ebi.ac.uk/ols/index)
-- [BioPortal](https://bioportal.bioontology.org/)
+- [ChEBI](https://www.ebi.ac.uk/chebi/)
+- [UniChem](https://www.ebi.ac.uk/unichem/)
+- [UniProt](https://www.uniprot.org/)
+- [RefMet](https://refmet.metabolomicsworkbench.org/)

{biomapper-0.3.1 → biomapper-0.3.2}/biomapper/__init__.py

@@ -3,5 +3,5 @@
 from .standardization import RaMPClient
 from .core import SetAnalyzer
 
-__version__ = "0.3.1"
+__version__ = "0.3.2"
 __all__ = ["RaMPClient", "SetAnalyzer"]

biomapper-0.3.2/biomapper/mapping/llm_mapper.py

@@ -0,0 +1,130 @@
+"""Base LLM integration for biomapper."""
+
+from typing import Dict, List, Optional, TYPE_CHECKING
+
+import time
+import os
+
+from langfuse import Langfuse  # type: ignore
+from openai import OpenAI
+
+if TYPE_CHECKING:
+    from openai.types.chat import ChatCompletionMessageParam
+
+from ..schemas.llm_schema import (
+    LLMMatch,
+    LLMMapperResult,
+    LLMMapperMetrics,
+    MatchConfidence,
+)
+
+
+class LLMMapper:
+    """Base class for LLM-based ontology mapping."""
+
+    def __init__(
+        self,
+        model: str = "gpt-4",
+        temperature: float = 0.0,
+        max_tokens: int = 1000,
+    ):
+        """Initialize the LLM mapper."""
+        self.model = model
+        self.temperature = temperature
+        self.max_tokens = max_tokens
+
+        # Initialize OpenAI client
+        self.client = OpenAI()
+
+        # Initialize Langfuse
+        self.langfuse = Langfuse(
+            public_key=os.getenv("LANGFUSE_PUBLIC_KEY"),
+            secret_key=os.getenv("LANGFUSE_SECRET_KEY"),
+            host=os.getenv("LANGFUSE_HOST", "https://cloud.langfuse.com"),
+        )
+
+    def _create_system_prompt(self) -> str:
+        """Create the system prompt for ontology mapping."""
+        return (
+            "You are an expert at mapping chemical and biological terms to ontologies. "
+            "Your task is to find the most appropriate ontology term for a given input. "
+            "Consider synonyms, related terms, and the hierarchical structure of the ontology."
+        )
+
+    def _estimate_cost(self, tokens_used: int) -> float:
+        """Estimate the cost of the API call in USD."""
+        # GPT-4 pricing (as of 2023)
+        return tokens_used * 0.00003  # $0.03 per 1000 tokens
+
+    def map_term(
+        self,
+        term: str,
+        target_ontology: Optional[str] = None,
+        metadata: Optional[Dict[str, str]] = None,
+    ) -> LLMMapperResult:
+        """Map a single term using the LLM.
+
+        Args:
+            term: Input term to map
+            target_ontology: Optional target ontology identifier
+            metadata: Optional metadata to include
+
+        Returns:
+            LLMMapperResult containing matches and metrics
+        """
+        start_time = time.time()
+
+        # Create messages
+        messages: List[ChatCompletionMessageParam] = [
+            {"role": "system", "content": self._create_system_prompt()},
+            {"role": "user", "content": f"Map the following term: {term}"},
+        ]
+
+        if target_ontology:
+            messages.append(
+                {"role": "user", "content": f"Target ontology: {target_ontology}"}
+            )
+
+        # Make API call
+        response = self.client.chat.completions.create(
+            model=self.model,
+            messages=messages,
+            temperature=self.temperature,
+            max_tokens=self.max_tokens,
+        )
+
+        # Calculate metrics
+        end_time = time.time()
+        latency = (end_time - start_time) * 1000  # Convert to ms
+        tokens_used = response.usage.total_tokens if response.usage else 0
+
+        # Process response into matches
+        content = response.choices[0].message.content if response.choices else None
+        if not content:
+            raise ValueError("No content in LLM response")
+
+        matches = [
+            LLMMatch(
+                target_id="example_id",
+                target_name=content,
+                confidence=MatchConfidence.MEDIUM,
+                score=0.8,
+                reasoning="Based on LLM response",
+                metadata=metadata or {},
+            )
+        ]
+
+        # Create result
+        return LLMMapperResult(
+            query_term=term,
+            matches=matches,
+            best_match=matches[0] if matches else None,
+            metrics=LLMMapperMetrics(
+                latency_ms=latency,
+                tokens_used=tokens_used,
+                provider="openai",
+                model=self.model,
+                cost=self._estimate_cost(tokens_used),
+            ),
+            trace_id=self.langfuse.trace(name="llm_mapping").id,
+        )

{biomapper-0.3.1 → biomapper-0.3.2}/biomapper/mapping/metabolite_name_mapper.py

@@ -95,7 +95,9 @@ class MetaboliteNameMapper:
                 if inchikey:
                     try:
                         unichem_result = (
-                            self.unichem_client.get_compound_info_by_inchikey(inchikey)
+                            self.unichem_client.get_compound_info_by_src_id(
+                                inchikey, "inchikey"
+                            )
                         )
                         if unichem_result:
                             # Get ChEBI ID if not already found