satif-ai 0.2.3__py3-none-any.whl → 0.2.5__py3-none-any.whl

satif_ai/code_builders/transformation.py

@@ -1,10 +1,11 @@
+import base64
 import os
 import re
 from pathlib import Path
 from typing import Dict, List, Optional, Union
 
 from agents import Agent, Runner, function_tool
-from agents.mcp.server import MCPServerStdio
+from agents.mcp.server import MCPServer
 from mcp import ClientSession
 from satif_core import AsyncCodeBuilder, CodeBuilder, SDIFDatabase
 from satif_sdk.comparators import get_comparator
@@ -14,6 +15,7 @@ from satif_sdk.transformers import CodeTransformer
 # Global variables for transformation
 INPUT_SDIF_PATH: Optional[Path] = None
 OUTPUT_TARGET_FILES: Optional[Dict[Union[str, Path], str]] = None
+SCHEMA_ONLY: Optional[bool] = None
 
 
 @function_tool
@@ -31,6 +33,9 @@ async def execute_transformation(code: str) -> str:
     generated_output_path = code_transformer.export(INPUT_SDIF_PATH)
 
     comparisons = []
+    comparator_kwargs = {}
+    if SCHEMA_ONLY:
+        comparator_kwargs["check_structure_only"] = True
 
     if os.path.isdir(generated_output_path):
         # If it's a directory, compare each file with its corresponding target
@@ -45,7 +50,9 @@ async def execute_transformation(code: str) -> str:
                     generated_output_path, output_target_file_name
                 )
                 comparator = get_comparator(output_target_file_name.split(".")[-1])
-                comparison = comparator.compare(generated_file_path, output_base_file)
+                comparison = comparator.compare(
+                    generated_file_path, output_base_file, **comparator_kwargs
+                )
                 comparisons.append(
                     f"Comparison for {generated_file_path} [SOURCE] with {output_target_file_name} [TARGET]: {comparison}"
                 )
@@ -59,7 +66,9 @@ async def execute_transformation(code: str) -> str:
             output_file = list(OUTPUT_TARGET_FILES.keys())[0]
             output_target_file_name = list(OUTPUT_TARGET_FILES.values())[0]
             comparator = get_comparator(output_file.split(".")[-1])
-            comparison = comparator.compare(generated_output_path, output_file)
+            comparison = comparator.compare(
+                generated_output_path, output_file, **comparator_kwargs
+            )
             comparisons.append(
                 f"Comparison for {generated_output_path} [SOURCE] with {output_target_file_name} [TARGET]: {comparison}"
             )
@@ -88,9 +97,9 @@ class TransformationAsyncCodeBuilder(AsyncCodeBuilder):
 
     def __init__(
         self,
-        mcp_server: MCPServerStdio,
+        mcp_server: MCPServer,
         mcp_session: ClientSession,
-        llm_model: str = "o3-mini",
+        llm_model: str = "o4-mini",
     ):
         self.mcp_server = mcp_server
         self.mcp_session = mcp_session
@@ -98,39 +107,116 @@ class TransformationAsyncCodeBuilder(AsyncCodeBuilder):
 
     async def build(
         self,
-        sdif: Path,
+        sdif: Path,  # This will now be relative to project root (MCP server CWD)
         output_target_files: Dict[Union[str, Path], str] | List[Path],
-        output_sdif: Optional[Path] = None,
+        output_sdif: Optional[Path] = None,  # This will now be relative or None
         instructions: Optional[str] = None,
+        schema_only: bool = False,
     ) -> str:
-        global INPUT_SDIF_PATH, OUTPUT_TARGET_FILES
-        INPUT_SDIF_PATH = Path(sdif)
+        global INPUT_SDIF_PATH, OUTPUT_TARGET_FILES, SCHEMA_ONLY
+        # INPUT_SDIF_PATH is used by execute_transformation tool, needs to be accessible from where that tool runs.
+        # If execute_transformation runs in the same process as the builder, absolute path is fine.
+        # If it were a separate context, this might need adjustment.
+        # For now, assume execute_transformation can access absolute paths if needed for its *input SDIF*.
+        # However, the sdif for MCP URIs must be relative.
+        INPUT_SDIF_PATH = Path(sdif).resolve()
+        SCHEMA_ONLY = schema_only
+        # Paths for MCP URIs are now expected to be relative to MCP server CWD (project root)
+        # So, use them directly as strings.
+        input_sdif_mcp_uri_path = base64.b64encode(str(sdif).encode()).decode()
+        output_sdif_mcp_uri_path = (
+            base64.b64encode(str(output_sdif).encode()).decode()
+            if output_sdif
+            else None
+        )
 
-        if isinstance(output_target_files, list):
-            OUTPUT_TARGET_FILES = {file: file.name for file in output_target_files}
-        else:
-            OUTPUT_TARGET_FILES = output_target_files
+        input_schema = await self.mcp_session.read_resource(
+            f"schema://{input_sdif_mcp_uri_path}"
+        )
+        input_sample = await self.mcp_session.read_resource(
+            f"sample://{input_sdif_mcp_uri_path}"
+        )
 
-        input_schema = await self.mcp_session.read_resource(f"schema://{sdif}")
-        input_sample = await self.mcp_session.read_resource(f"sample://{sdif}")
+        output_schema_text = "N/A"
+        output_sample_text = "N/A"
+        if output_sdif_mcp_uri_path:
+            try:
+                output_schema_content = await self.mcp_session.read_resource(
+                    f"schema://{output_sdif_mcp_uri_path}"
+                )
+                if output_schema_content.contents:
+                    output_schema_text = output_schema_content.contents[0].text
+            except Exception as e:
+                print(
+                    f"Warning: Could not read schema for output_sdif {output_sdif_mcp_uri_path}: {e}"
+                )
 
-        output_schema = await self.mcp_session.read_resource(f"schema://{output_sdif}")
-        output_sample = await self.mcp_session.read_resource(f"sample://{output_sdif}")
-        output_representation = {
-            file: get_representer(file).represent(file)
-            for file in list(OUTPUT_TARGET_FILES.keys())
-        }
+            try:
+                output_sample_content = await self.mcp_session.read_resource(
+                    f"sample://{output_sdif_mcp_uri_path}"
+                )
+                if output_sample_content.contents:
+                    output_sample_text = output_sample_content.contents[0].text
+            except Exception as e:
+                print(
+                    f"Warning: Could not read sample for output_sdif {output_sdif_mcp_uri_path}: {e}"
+                )
+
+        # OUTPUT_TARGET_FILES keys are absolute paths to original example files for local reading by representers/comparators.
+        # Values are agent-facing filenames.
+        if isinstance(output_target_files, list):
+            OUTPUT_TARGET_FILES = {
+                file_path.resolve(): file_path.name for file_path in output_target_files
+            }
+        elif isinstance(output_target_files, dict):
+            temp_map = {}
+            for k, v in output_target_files.items():
+                if isinstance(k, Path):
+                    temp_map[k.resolve()] = v
+                else:
+                    temp_map[k] = v
+            OUTPUT_TARGET_FILES = temp_map
+        else:
+            OUTPUT_TARGET_FILES = {}
+
+        output_representation = {}
+        if OUTPUT_TARGET_FILES:
+            for file_key_abs_path in list(OUTPUT_TARGET_FILES.keys()):
+                agent_facing_name = OUTPUT_TARGET_FILES[file_key_abs_path]
+                print(f"Representing {agent_facing_name} from {file_key_abs_path}!!")
+                try:
+                    # Representer uses the absolute path (file_key_abs_path) to read the example file.
+                    output_representation[agent_facing_name] = get_representer(
+                        file_key_abs_path
+                    ).represent(file_key_abs_path)
+                except Exception as e:
+                    print(
+                        f"Warning: Could not get representation for {agent_facing_name} (path {file_key_abs_path}): {e}"
+                    )
+                    output_representation[agent_facing_name] = (
+                        f"Error representing file: {e}"
+                    )
 
         prompt = await self.mcp_session.get_prompt(
             "create_transformation",
             arguments={
-                "input_file": INPUT_SDIF_PATH.name,
-                "input_schema": input_schema.contents[0].text,
-                "input_sample": input_sample.contents[0].text,
+                "input_file": Path(
+                    input_sdif_mcp_uri_path
+                ).name,  # Display name for prompt (from relative path)
+                "input_schema": input_schema.contents[0].text
+                if input_schema.contents
+                else "Error reading input schema",
+                "input_sample": input_sample.contents[0].text
+                if input_sample.contents
+                else "Error reading input sample",
                 "output_files": str(list(OUTPUT_TARGET_FILES.values())),
-                "output_schema": output_schema.contents[0].text,
-                "output_sample": output_sample.contents[0].text,
-                "output_representation": str(output_representation),
+                "output_schema": output_schema_text,
+                "output_sample": output_sample_text
+                if not SCHEMA_ONLY
+                else "Sample not available. Transform according to the schema only. No data is present in the output example.",
+                "output_representation": str(
+                    output_representation
+                ),  # Representation keyed by agent-facing name
             },
         )
         agent = Agent(

{satif_ai-0.2.3.dist-info → satif_ai-0.2.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: satif-ai
-Version: 0.2.3
+Version: 0.2.5
 Summary: AI Agents for Satif
 License: MIT
 Author: Bryan Djafer

{satif_ai-0.2.3.dist-info → satif_ai-0.2.5.dist-info}/RECORD

@@ -3,15 +3,15 @@ satif_ai/adapters/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU
 satif_ai/adapters/tidy.py,sha256=2oYj7Gz3vOQtzcpoJI4JbftWlMKvOWL8rdwthjg-zUE,19884
 satif_ai/code_builders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 satif_ai/code_builders/adaptation.py,sha256=E29YM0S6pMtAfB0uzSUexoeWKwXfF8iJVyYUCKWQz5k,188
-satif_ai/code_builders/transformation.py,sha256=fdegXe2jNYxbPWOoDKhpKws8ltO5we-aJQQAhBBYVWQ,5970
+satif_ai/code_builders/transformation.py,sha256=ZYpZpJb28PzN3suNG1lxeF2zB6mWn8gMoOLQov9YJUA,10043
 satif_ai/plot_builders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 satif_ai/plot_builders/agent.py,sha256=Ncw7SL9qkpRN0hw76ezSo1K8vVQK6gcXFp8x8VFwqUI,8291
 satif_ai/plot_builders/prompt.py,sha256=m0W1SsxnB9_FhIYRumkthImJbK-7KUm4dygN3kjAXGk,6877
 satif_ai/plot_builders/tool.py,sha256=MeLnG_wFoITSVWZcNFsQLCi157O4L3wItQgolBa4fAw,5994
 satif_ai/standardizers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 satif_ai/standardizers/ai_csv.py,sha256=AAeTt7eqFAtayxF2b95Z_K_lnMdwBBnv2Cn-qTEpMp8,29499
-satif_ai-0.2.3.dist-info/LICENSE,sha256=kS8EN6yAaGZd7V5z6GKSn_x3ozcZltrfRky4vMPRCw8,1072
-satif_ai-0.2.3.dist-info/METADATA,sha256=lPxjO5D5g8U-kePulS6-15OsTMkosvt49xn6KX1oiWI,670
-satif_ai-0.2.3.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-satif_ai-0.2.3.dist-info/entry_points.txt,sha256=_8RP7sSos2cwqYblhwHOK3hhRkLbP6XkbAju3vHWft8,40
-satif_ai-0.2.3.dist-info/RECORD,,
+satif_ai-0.2.5.dist-info/LICENSE,sha256=kS8EN6yAaGZd7V5z6GKSn_x3ozcZltrfRky4vMPRCw8,1072
+satif_ai-0.2.5.dist-info/METADATA,sha256=jIyyKOcveIz_98_JTyF6bsfVn5iOVvc3qLFVQZ0wUSo,670
+satif_ai-0.2.5.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+satif_ai-0.2.5.dist-info/entry_points.txt,sha256=Mz2SwYALjktap1bF-Q3EWBgiZVNT6QJCVsCs_fCV33Y,43
+satif_ai-0.2.5.dist-info/RECORD,,

satif_ai-0.2.5.dist-info/entry_points.txt

@@ -0,0 +1,3 @@
+[console_scripts]
+satif-ai=satif.cli:main
+

satif_ai-0.2.3.dist-info/entry_points.txt

@@ -1,3 +0,0 @@
-[console_scripts]
-satif=satif.cli:main
-