aurelian 0.3.4__tar.gz → 0.4.1__tar.gz

{aurelian-0.3.4 → aurelian-0.4.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: aurelian
-Version: 0.3.4
+Version: 0.4.1
 Summary: aurelian
 License: MIT
 Author: Author 1
@@ -17,6 +17,7 @@ Provides-Extra: gocam
 Provides-Extra: gradio
 Provides-Extra: linkml
 Provides-Extra: pdfminer
+Provides-Extra: pypaperbot
 Provides-Extra: rdkit
 Requires-Dist: bioservices (>=1.12.0) ; extra == "bioservices"
 Requires-Dist: bs4 (>=0.0.2)
@@ -30,13 +31,14 @@ Requires-Dist: linkml ; extra == "linkml"
 Requires-Dist: linkml-store[llm,mongodb] (>=0.2.4)
 Requires-Dist: logfire (>=3.5.3)
 Requires-Dist: markdownify (>=0.14.1)
-Requires-Dist: markitdown (>=0.0.1a4)
+Requires-Dist: markitdown (>=0.1.1)
 Requires-Dist: mcp[cli] (>=1.3.0,<2.0.0)
 Requires-Dist: oaklib (>=0.6.19)
+Requires-Dist: onnxruntime (<=1.19.0)
 Requires-Dist: paper-qa (>=5.20.0,<6.0.0)
 Requires-Dist: pdfminer-six ; extra == "pdfminer"
-Requires-Dist: pydantic-ai (>=0.0.29)
-Requires-Dist: pypaperbot (>=1.4.1)
+Requires-Dist: pydantic-ai (>=0.2.0)
+Requires-Dist: pypaperbot (>=1.4.1) ; extra == "pypaperbot"
 Requires-Dist: pytest-asyncio (>=0.25.3,<0.26.0)
 Requires-Dist: rdkit ; extra == "rdkit" or extra == "chem"
 Requires-Dist: tabulate (>=0.9.0)

{aurelian-0.3.4 → aurelian-0.4.1}/pyproject.toml

@@ -1,6 +1,9 @@
+[project]
+name = "aurelian"
+
 [tool.poetry]
 name = "aurelian"
-version = "0.3.4"
+version = "0.4.1"
 description = "aurelian"
 authors = ["Author 1 <author@org.org>"]
 license = "MIT"
@@ -17,14 +20,15 @@ rdkit = {optional=true, version="*"}
 gocam = {optional=true, version="*"}
 linkml = {optional=true, version="*"}
 gradio = "^5.23.3"
-pydantic-ai = ">=0.0.29"
+pydantic-ai = ">=0.2.0"
 logfire = ">=3.5.3"
 bs4 = ">=0.0.2"
 duckduckgo-search = ">=7.3.2"
 markdownify = ">=0.14.1"
 wikipedia = ">=1.4.0"
-markitdown = ">=0.0.1a4"
-pypaperbot = ">=1.4.1"
+onnxruntime = "<=1.19.0"
+markitdown = ">=0.1.1"
+pypaperbot = {optional=true, version=">=1.4.1"}
 undetected-chromedriver = ">=3.5.5"
 bioservices = ">=1.12.0"
 pdfminer-six = {optional=true, version="*"}
@@ -54,6 +58,7 @@ bioservices = ["bioservices"]
 linkml = ["linkml"]
 chem = ["rdkit"]
 pdfminer = ["pdfminer-six"]
+pypaperbot = ["pypaperbot"]
 
 [tool.pytest.ini_options]
 markers = [
@@ -78,6 +83,7 @@ typer = ">=0.9.0"  # Required for mkdocs-click
 
 [tool.poetry.scripts]
 aurelian = "aurelian.cli:main"
+aurelian-mcp-gocam = "aurelian.agents.gocam.gocam_mcp:main"
 
 [tool.poetry-dynamic-versioning]
 enable = false

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/biblio/biblio_agent.py

@@ -30,6 +30,7 @@ biblio_agent = Agent(
         " tables for this. Use your judgment in how to roll up tables, and whether values"
         " should be present/absent, increased/decreased, or more specific."
     ),
+    defer_model_check=True,
 )
 
 

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/checklist/checklist_agent.py

@@ -22,6 +22,7 @@ checklist_agent = Agent(
         "\nThe available checklists are:"
     ),
     deps_type=ChecklistDependencies,
+    defer_model_check=True,
 )
 
 

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/chemistry/chemistry_agent.py

@@ -39,7 +39,8 @@ chemistry_agent = Agent(
         Tool(retrieve_chemistry_web_page),
         Tool(inspect_file),
         Tool(list_files),
-    ]
+    ],
+    defer_model_check=True,
 )
 
 # Remove the chat import to avoid circular imports

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/d4d/__init__.py

@@ -3,7 +3,7 @@ D4D (Datasheets for Datasets) agent package for extracting dataset metadata.
 """
 
 # isort: skip_file
-from .d4d_agent import data_sheets_agent  # noqa: E402
+from .d4d_agent import d4d_agent  # noqa: E402
 from .d4d_config import D4DConfig, get_config  # noqa: E402
 from .d4d_gradio import chat  # noqa: E402
 from .d4d_tools import (  # noqa: E402
@@ -14,7 +14,7 @@ from .d4d_tools import (  # noqa: E402
 
 __all__ = [
     # Agent
-    "data_sheets_agent",
+    "d4d_agent",
     
     # Config
     "D4DConfig",

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/d4d/d4d_agent.py

@@ -8,7 +8,7 @@ from .d4d_tools import get_full_schema, process_website_or_pdf
 
 
 # Create the agent, the full schema will be loaded when needed
-data_sheets_agent = Agent(
+d4d_agent = Agent(
     model="openai:gpt-4o",
     deps_type=D4DConfig,
     system_prompt="""
@@ -21,10 +21,11 @@ content, extract all the relevant metadata, and output a YAML document that exac
 conforms to the above schema. The output must be valid YAML with all required fields 
 filled in, following the schema exactly.
 """,
+    defer_model_check=True,
 )
 
 
-@data_sheets_agent.system_prompt
+@d4d_agent.system_prompt
 async def add_schema(ctx: RunContext[D4DConfig]) -> str:
     """
     Add the full schema to the system prompt.
@@ -39,7 +40,7 @@ async def add_schema(ctx: RunContext[D4DConfig]) -> str:
     return schema
 
 
-@data_sheets_agent.tool
+@d4d_agent.tool
 async def extract_metadata(ctx: RunContext[D4DConfig], url: str) -> str:
     """
     Extract metadata from a dataset description document or webpage.

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/d4d/d4d_gradio.py

@@ -5,7 +5,7 @@ from typing import List, Optional
 
 import gradio as gr
 
-from .d4d_agent import data_sheets_agent
+from .d4d_agent import d4d_agent
 from .d4d_config import D4DConfig, get_config
 
 
@@ -22,7 +22,7 @@ async def process_url(url: str, history: List[str], config: D4DConfig) -> str:
         YAML formatted metadata
     """
     # Run the agent with the URL
-    result = await data_sheets_agent.run(url, deps=config)
+    result = await d4d_agent.run(url, deps=config)
     return result.data
 
 

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/diagnosis/diagnosis_agent.py

@@ -44,6 +44,7 @@ diagnosis_agent = Agent(
     deps_type=DiagnosisDependencies,
     result_type=str,
     system_prompt=DIAGNOSIS_SYSTEM_PROMPT,
+    defer_model_check=True,
 )
 
 # Register tools

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/gocam/__init__.py

@@ -3,11 +3,13 @@ GOCAM agent module for working with Gene Ontology Causal Activity Models.
 """
 from pathlib import Path
 
+from ...evaluators.substring_evaluator import SubstringEvaluator
+
 THIS_DIR = Path(__file__).parent
 DOCUMENTS_DIR = THIS_DIR / "documents"
 
 # isort: skip_file
-from .gocam_agent import gocam_agent  # noqa: E402
+from .gocam_agent import gocam_agent, gocam_reviewer_agent, gocam_review_summarizer_agent  # noqa: E402
 from .gocam_config import GOCAMDependencies, get_config  # noqa: E402
 from .gocam_gradio import chat  # noqa: E402
 from .gocam_tools import (  # noqa: E402
@@ -18,6 +20,7 @@ from .gocam_tools import (  # noqa: E402
     fetch_document,
     validate_gocam_model,
 )
+from .gocam_evals import create_eval_dataset  # noqa: E402
 
 __all__ = [
     # Constants
@@ -26,6 +29,8 @@ __all__ = [
     
     # Agent
     "gocam_agent",
+    "gocam_reviewer_agent",
+    "gocam_review_summarizer_agent",
     # Config
     "GOCAMDependencies",
     "get_config",
@@ -40,4 +45,8 @@ __all__ = [
     
     # Gradio
     "chat",
+    
+    # Evals
+    "create_eval_dataset",
+    "SubstringEvaluator",
 ]

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/gocam/gocam_agent.py

@@ -135,6 +135,7 @@ gocam_agent = Agent(
     deps_type=GOCAMDependencies,
     system_prompt=SYSTEM,
     tools=core_tools,
+    defer_model_check=True,
 )
 
 def get_documents_for_prompt() -> str:
@@ -208,6 +209,7 @@ gocam_reviewer_agent = Agent(
         Tool(lookup_gocam_local),
         #Tool(validate_gocam_model),
     ],
+    defer_model_check=True,
 )
 
 
@@ -237,4 +239,5 @@ gocam_review_summarizer_agent = Agent(
         #Tool(lookup_gocam_local),
     ],
     result_type=GOCamReviewSummary,
+    defer_model_check=True,
 )

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/gocam/gocam_evals.py

@@ -14,9 +14,6 @@ from pydantic_evals import Case, Dataset
 from aurelian.agents.gocam.gocam_agent import gocam_agent
 from aurelian.agents.gocam.gocam_config import GOCAMDependencies
 
-class GOCAMMetadata(Dict[str, Any]):
-    """Simple metadata dictionary for GO-CAM evaluations."""
-    pass
 
 # Define individual evaluation cases
 case1 = Case(

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/linkml/linkml_mcp.py

@@ -56,7 +56,6 @@ async def inspect_file(data_file: str) -> str:
     Inspect a file in the working directory.
 
     Args:
-        ctx:
         data_file: name of file
 
     Returns:
@@ -71,7 +70,6 @@ async def list_files() -> str:
     List files in the working directory.
 
     Args:
-        ctx:
 
     Returns:
 
@@ -84,7 +82,6 @@ async def write_to_file(data: str, file_name: str) -> str:
     Write data to a file in the working directory.
 
     Args:
-        ctx:
         data:
         file_name:
 
@@ -104,7 +101,6 @@ async def validate_data(schema: str, data_file: str) -> str:
     You can write data to the working directory using the `write_to_file` tool.
 
     Args:
-        ctx:
         schema: the schema (as a YAML string)
         data_file: the name of the data file in the working directory
 
@@ -167,7 +163,6 @@ async def download_web_page(url: str, local_file_name: str) -> str:
     Download contents of a web page.
 
     Args:
-        ctx:
         url: URL of the web page
         local_file_name: Name of the local file to save the
 
@@ -183,4 +178,4 @@ async def download_web_page(url: str, local_file_name: str) -> str:
 
 if __name__ == "__main__":
     # Initialize and run the server
-    mcp.run(transport='stdio')
+    mcp.run(transport='stdio')

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/literature/literature_agent.py

@@ -1,6 +1,7 @@
 """
 Agent for working with scientific literature and publications.
 """
+from aurelian.agents.web.web_tools import perplexity_query
 from aurelian.agents.literature.literature_config import LiteratureDependencies
 from aurelian.agents.literature.literature_tools import (
     lookup_pmid,
@@ -48,6 +49,25 @@ literature_agent = Agent(
         Tool(get_article_abstract),
         Tool(extract_text_from_pdf_url),
         Tool(search_literature_web),
+        #Tool(perplexity_query), 
+        Tool(retrieve_literature_page),
+        Tool(inspect_file),
+        Tool(list_files),
+    ]
+)
+
+advanced_literature_agent = Agent(
+    model="openai:gpt-4o",
+    deps_type=LiteratureDependencies,
+    system_prompt=SYSTEM,
+    tools=[
+        Tool(lookup_pmid),
+        Tool(lookup_doi),
+        Tool(convert_pmid_to_doi),
+        Tool(convert_doi_to_pmid),
+        Tool(get_article_abstract),
+        Tool(extract_text_from_pdf_url),
+        Tool(perplexity_query), 
         Tool(retrieve_literature_page),
         Tool(inspect_file),
         Tool(list_files),

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/monarch/monarch_agent.py

@@ -37,6 +37,7 @@ monarch_agent = Agent(
     model="openai:gpt-4o",
     system_prompt=MONARCH_SYSTEM_PROMPT,
     deps_type=MonarchDependencies,
+    defer_model_check=True,
 )
 
 # Register the tools with the agent

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/monarch/monarch_tools.py

@@ -7,7 +7,6 @@ from typing import Dict, List, Optional
 from pydantic_ai import RunContext, ModelRetry
 
 from aurelian.utils.data_utils import obj_to_dict
-from aurelian.utils.async_utils import run_sync
 from .monarch_config import MonarchDependencies, get_config
 
 

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/ontology_mapper/ontology_mapper_agent.py

@@ -39,6 +39,7 @@ ontology_mapper_agent = Agent(
     deps_type=OntologyMapperDependencies,
     result_type=str,
     system_prompt=ONTOLOGY_MAPPER_SYSTEM_PROMPT,
+    defer_model_check=True,
 )
 
 # Register the tools with the agent

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/paperqa/paperqa_agent.py

@@ -55,6 +55,7 @@ paperqa_agent = Agent(
     deps_type=PaperQADependencies,
     result_type=str,
     system_prompt=PAPERQA_SYSTEM_PROMPT,
+    defer_model_check=True,
 )
 
 paperqa_agent.tool(search_papers)

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/rag/rag_agent.py

@@ -18,6 +18,7 @@ rag_agent = Agent(
         "  - `inspect_document` to retrieve a specific document (by title/name)"
         "You can also use `lookup_pmid` to retrieve the text of a PubMed ID, or `search_web` to search the web."
     ),
+    defer_model_check=True,
 )
 
 

aurelian-0.4.1/src/aurelian/agents/talisman/talisman_mcp.py

@@ -0,0 +1,75 @@
+"""
+MCP tools for retrieving gene information using the UniProt API and NCBI Entrez.
+"""
+import os
+
+from mcp.server.fastmcp import FastMCP
+
+from aurelian.agents.talisman.talisman_agent import TALISMAN_SYSTEM_PROMPT
+from aurelian.agents.talisman.talisman_config import TalismanConfig, get_config
+from aurelian.agents.talisman.talisman_tools import (
+    get_gene_description,
+    get_gene_descriptions,
+    get_genes_from_list,
+    analyze_gene_set
+)
+
+from pydantic_ai import RunContext
+
+# Initialize FastMCP server
+mcp = FastMCP("talisman", instructions=TALISMAN_SYSTEM_PROMPT)
+
+def deps() -> TalismanConfig:
+    """Get the Talisman dependencies."""
+    return get_config()
+
+def ctx() -> RunContext[TalismanConfig]:
+    """Get the run context with dependencies."""
+    rc: RunContext[TalismanConfig] = RunContext[TalismanConfig](
+        deps=deps(),
+        model=None, usage=None, prompt=None,
+    )
+    return rc
+
+@mcp.tool()
+async def get_gene_info(gene_id: str) -> str:
+    """
+    Get description for a single gene ID.
+    
+    Args:
+        gene_id: The gene identifier (UniProt ID, gene symbol, etc.)
+        
+    Returns:
+        The gene description in a structured format
+    """
+    return get_gene_description(ctx(), gene_id)
+
+@mcp.tool()
+async def get_multiple_gene_info(gene_ids: str) -> str:
+    """
+    Get descriptions for multiple gene IDs provided as a string.
+    
+    Args:
+        gene_ids: String containing gene identifiers separated by commas, spaces, or newlines
+        
+    Returns:
+        The gene descriptions in a structured format
+    """
+    return get_genes_from_list(ctx(), gene_ids)
+
+@mcp.tool()
+async def analyze_genes(gene_list: str) -> str:
+    """
+    Analyze a set of genes and generate a biological summary of their properties and relationships.
+    
+    Args:
+        gene_list: String containing gene identifiers separated by commas, spaces, or newlines
+        
+    Returns:
+        A structured biological summary of the gene set with Narrative, Functional Terms Table, and Gene Summary Table
+    """
+    return analyze_gene_set(ctx(), gene_list)
+
+if __name__ == "__main__":
+    # Initialize and run the server
+    mcp.run(transport='stdio')

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/ubergraph/ubergraph_agent.py

@@ -50,6 +50,7 @@ ubergraph_agent = Agent(
     "openai:gpt-4o",
     deps_type=Dependencies,
     result_type=str,
+    defer_model_check=True,
 )
 
 # Register tools

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/agents/web/web_tools.py

@@ -43,9 +43,11 @@ async def search_web(query: str) -> str:
     return web_search(query)
 
 
-async def perplexity_query(query: str, model_name: str = "sonar") -> ResultWithCitations:
+async def perplexity_query(query: str, model_name: str = "sonar-pro") -> ResultWithCitations:
     """
     Query the Perplexity API and return structured results with citations.
+
+    The Perplexity API performs a web search and returns a structured response with citations.
     
     Args:
         query: The query to send to Perplexity
@@ -58,14 +60,25 @@ async def perplexity_query(query: str, model_name: str = "sonar") -> ResultWithC
         ValueError: If the Perplexity API key is not set
         RuntimeError: If the response parsing fails
     """
+    # TODO: consider using perplexity API directly, gives control over search domains, e.g. https://docs.perplexity.ai/guides/search-domain-filters
     perplexity_api_key = os.environ.get("PERPLEXITY_API_KEY")
     if not perplexity_api_key:
         raise ValueError("PERPLEXITY_API_KEY environment variable is not set")
         
+    # Use specific implementation instead of OpenAIModel directly
+    # since OpenAIModel doesn't accept base_url and api_key params directly
+    from pydantic_ai.models.openai import Provider
+    from pydantic_ai.providers.openai import OpenAIProvider
+
+    
+    provider = OpenAIProvider(
+        api_key=perplexity_api_key,
+        base_url='https://api.perplexity.ai'
+    )
+    
     sonar_model = OpenAIModel(
         model_name=model_name,
-        base_url='https://api.perplexity.ai',
-        api_key=perplexity_api_key,
+        provider=provider,
     )
     
     agent = Agent(sonar_model,

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/cli.py

@@ -4,6 +4,7 @@ import logging
 import os
 from typing import Any, Awaitable, Callable, Optional, List
 
+from aurelian.utils.async_utils import run_sync
 import click
 from pydantic_ai.models.openai import OpenAIModel
 from pydantic_ai.providers.openai import OpenAIProvider
@@ -550,6 +551,14 @@ def websearch(term):
     txt = web_search(term)
     print(txt)
 
+@main.command()
+@click.argument("term")
+def perplexity(term):
+    """Search the web for a query term, with citations."""
+    from aurelian.agents.web.web_tools import perplexity_query
+    result = run_sync(perplexity_query(term))
+    import yaml
+    print(yaml.dump(result.model_dump(), indent=2))
 
 @main.command()
 @click.argument("url")

aurelian-0.4.1/src/aurelian/evaluators/model.py

@@ -0,0 +1,9 @@
+# Use TypedDict for metadata
+from typing_extensions import TypedDict
+
+
+MetadataDict = TypedDict("Metadata", {"difficulty": str, "type": str})
+
+
+def metadata(difficulty: str, type: str) -> MetadataDict:
+    return {"difficulty": difficulty, "type": type}

aurelian-0.4.1/src/aurelian/evaluators/substring_evaluator.py

@@ -0,0 +1,30 @@
+from pydantic_evals.evaluators import Evaluator, EvaluatorContext
+
+
+class SubstringEvaluator(Evaluator[str, str]):
+    """
+    Custom evaluator for GO-CAM agent responses.
+
+    This evaluator checks if the expected substring is present in the agent's response.
+    If no expected output is specified, it assumes the test passes (for cases where we
+    only want to verify the agent doesn't error out).
+    """
+
+    def evaluate(self, ctx: EvaluatorContext[str, str]) -> float:
+        """
+        Evaluate GO-CAM agent response by checking for expected substring.
+
+        Args:
+            ctx: The evaluator context containing input, output, and expected output
+
+        Returns:
+            Score between 0.0 and 1.0 (1.0 = pass, 0.0 = fail)
+        """
+        # If no expected output is specified, return 1.0 (success)
+        if ctx.expected_output is None:
+            return 1.0
+
+        # Check if expected string is in output
+        if ctx.expected_output.lower() in ctx.output.lower():
+            return 1.0
+        return 0.0

{aurelian-0.3.4 → aurelian-0.4.1}/src/aurelian/utils/async_utils.py

@@ -1,11 +1,14 @@
 import asyncio
-from typing import Callable
+from typing import Callable, Coroutine
 
 
-def run_sync(f: Callable):
+def run_sync(f: Callable | Coroutine):
     loop = asyncio.new_event_loop()
     asyncio.set_event_loop(loop)
-    result = f()
+    if isinstance(f, Coroutine):
+        result = f
+    else:
+        result = f()
 
     # Ensure it's a coroutine before running it
     if asyncio.iscoroutine(result):

aurelian-0.3.4/src/aurelian/agents/monarch/__init__.py

@@ -1,25 +0,0 @@
-"""
-Monarch agent package for interacting with the Monarch Knowledge Base.
-"""
-
-from .monarch_agent import monarch_agent, MONARCH_SYSTEM_PROMPT
-from .monarch_config import MonarchDependencies, get_config
-from .monarch_gradio import chat
-from .monarch_tools import find_gene_associations, find_disease_associations
-
-__all__ = [
-    # Agent
-    "monarch_agent",
-    "MONARCH_SYSTEM_PROMPT",
-    
-    # Config
-    "MonarchDependencies",
-    "get_config",
-    
-    # Tools
-    "find_gene_associations",
-    "find_disease_associations",
-    
-    # Gradio
-    "chat",
-]