Misc changes (#18)

Author: ludomitch

pyproject.toml

@@ -13,20 +13,15 @@ dependencies = [
   "pandas==2.2.3",
   "numpy==2.2.3",
   "matplotlib==3.10.0",
-  "scipy==1.15.2",
-  "seaborn==0.13.2",
-  "scikit-learn==1.6.1",
-  "statsmodels==0.14.4",
   "aiofiles==24.1.0",
   "google-auth==2.38.0",
   "google-cloud-storage==3.0.0",
   "google-cloud-secret-manager==2.23.0",
-  "futurehouse-client==0.3.15.dev71",
+  "futurehouse-client==0.3.18.dev25",
   "jupyter==1.1.1",
   "nbconvert==7.16.6",
   "notebook==7.3.2",
-  "nbformat==5.10.4",
-  "pydeseq2==0.5.0"
+  "nbformat==5.10.4"
 ]
 description = "Data analysis crow"
 name = "fhda"

src/fhda/config.py

@@ -25,4 +25,4 @@
 # FutureHosue client config
 ENVIRONMENT = os.getenv("ENVIRONMENT", "prod")
 CROW_STAGE = getattr(Stage, ENVIRONMENT.upper(), Stage.PROD)
-PLATFORM_API_KEY = os.getenv("CROW_API_KEY")
+PLATFORM_API_KEY = os.getenv("CROW_API_KEY", None)

src/fhda/data_analysis_env.py

@@ -96,6 +96,10 @@ async def query_literature(self, query: str) -> str:
         Args:
             query: The scientific question to answer
         """
+
+        if cfg.PLATFORM_API_KEY is None:
+            raise Exception("Platform API key is not set")
+
         logger.info("Running PQA query")
         client = FutureHouseClient(
             stage=cfg.CROW_STAGE,
@@ -197,27 +201,33 @@ def from_task(
             user_id = "default_user"
         if trajectory_id is None:
             trajectory_id = f"{gcs_artifact_path}-{time.time()}"
-
+        if environment_config:
+            kwargs = {
+                k: v
+                for k, v in environment_config.items()
+                if k in cfg.VALID_FROM_TASK_KWARGS
+            }
+        else:
+            kwargs = {}
+            environment_config = {}
         # Always create a new directory for the trajectory
         trajectory_path = (
             cfg.DATA_STORAGE_PATH / "user_trajectories" / user_id / trajectory_id
         )
+        if environment_config.get("gcs_override", False):
+            data_path = cfg.DATA_STORAGE_PATH / gcs_artifact_path
+        else:
+            data_path = (
+                cfg.DATA_STORAGE_PATH / "user_data" / user_id / gcs_artifact_path
+            )
         logger.info("Trajectory path: %s", trajectory_path)
+        logger.info("Data path: %s", data_path)
         trajectory_path.mkdir(parents=True, exist_ok=True)
-        for item in (cfg.DATA_STORAGE_PATH / gcs_artifact_path).iterdir():
+        for item in data_path.iterdir():
             if item.is_file():
                 shutil.copy2(item, trajectory_path)
             elif item.is_dir():
                 shutil.copytree(item, trajectory_path / item.name, dirs_exist_ok=True)
-        if environment_config:
-            kwargs = {
-                k: v
-                for k, v in environment_config.items()
-                if k in cfg.VALID_FROM_TASK_KWARGS
-            }
-        else:
-            kwargs = {}
-            environment_config = {}
         logger.info("Filtered kwargs: %s", kwargs)
 
         language = getattr(NBLanguage, environment_config.get("language", "PYTHON"))

src/fhda/prompts.py

@@ -40,12 +40,7 @@
    variable_name <- read_excel("<fpath>.csv", col_names = FALSE, .name_repair = "minimal")
    ```
 
-3. When printing dataframes, always wrap them in print() statements:
-   ```r
-   print(head(dataframe))
-   ```
-
-4. Very important: always use the tidyverse package where possible.
+3. Very important: always use the tidyverse package where possible.
 """
 
 

src/scripts/deploy.py

@@ -12,12 +12,12 @@
 )
 from futurehouse_client.models.app import TaskQueuesConfig
 
-HIGH = True
+HIGH = False
 ENVIRONMENT = "DEV"
 
 ENV_VARS = {
-    "OPENAI_API_KEY": os.environ["OPENAI_API_KEY"],
-    "ANTHROPIC_API_KEY": os.environ["ANTHROPIC_API_KEY"],
+    # "OPENAI_API_KEY": os.environ["OPENAI_API_KEY"],
+    # "ANTHROPIC_API_KEY": os.environ["ANTHROPIC_API_KEY"],
     "USE_DOCKER": "false",
     "STAGE": ENVIRONMENT,
     "ENVIRONMENT": ENVIRONMENT,
@@ -36,17 +36,85 @@
 TEMPERATURE = 1
 NUM_RETRIES = 3
 
-agent = AgentConfig(
-    agent_type="ReActAgent",
-    agent_kwargs={
+# agent = AgentConfig(
+#     agent_type="ReActAgent",
+#     agent_kwargs={
+#         "llm_model": {
+#             "name": MODEL,
+#             "temperature": TEMPERATURE,
+#             "num_retries": NUM_RETRIES,
+#         },
+#         "hide_old_env_states": True,
+#     },
+# )
+
+AGENT_MODEL_LIST = [
+    {
+        "model_name": "anthropic/claude-3-7-sonnet-20250219",
+        "litellm_params": {
+            "model": "anthropic/claude-3-7-sonnet-20250219",
+            "api_key": os.environ["ANTHROPIC_API_KEY"],
+        },
+    },
+    {
+        "model_name": "openai/gpt-4.1-2025-04-14",
+        "litellm_params": {
+            "model": "openai/gpt-4.1-2025-04-14",
+            "api_key": os.environ["OPENAI_API_KEY"],
+        },
+    },
+    {
+        "model_name": "anthropic/claude-3-5-sonnet-20241022",
+        "litellm_params": {
+            "model": "anthropic/claude-3-5-sonnet-20241022",
+            "api_key": os.environ["ANTHROPIC_API_KEY"],
+        },
+    },
+    {
+        "model_name": "openai/gpt-4o-2024-11-20",
+        "litellm_params": {
+            "model": "openai/gpt-4o-2024-11-20",
+            "api_key": os.environ["OPENAI_API_KEY"],
+        },
+    },
+]
+
+AGENT_ROUTER_KWARGS = {
+    "set_verbose": True,
+    # fallback in list order if the main key fails
+    "fallbacks": [
+        {
+            "openai/gpt-4.1-2025-04-14": [
+                "anthropic/claude-3-7-sonnet-20250219",
+                "anthropic/claude-3-5-sonnet-20241022",
+                "openai/gpt-4o-2024-11-20",
+            ]
+        }
+    ],
+}
+
+AGENT_CONFIG = {
+    "agent_type": "ReActAgent",
+    "agent_kwargs": {
         "llm_model": {
-            "name": MODEL,
-            "temperature": TEMPERATURE,
-            "num_retries": NUM_RETRIES,
+            "name": "anthropic/claude-3-7-sonnet-20250219",
+            "config": {
+                "model_list": AGENT_MODEL_LIST,
+                "router_kwargs": AGENT_ROUTER_KWARGS,
+                "fallbacks": [
+                    {
+                        "openai/gpt-4.1-2025-04-14": [
+                            "anthropic/claude-3-7-sonnet-20250219",
+                            "anthropic/claude-3-5-sonnet-20241022",
+                            "openai/gpt-4o-2024-11-20",
+                        ]
+                    }
+                ],
+            },
         },
         "hide_old_env_states": True,
     },
-)
+}
 
 CROWS_TO_DEPLOY = [
     JobDeploymentConfig(
@@ -55,8 +123,7 @@
         name="data-analysis-crow-high" if HIGH else "data-analysis-crow",
         environment="src.fhda.data_analysis_env.DataAnalysisEnv",
         environment_variables=ENV_VARS,
-        # agent="ldp.agent.ReActAgent",
-        agent=agent,
+        agent=AgentConfig(**AGENT_CONFIG),  # type: ignore
         container_config=CONTAINER_CONFIG,
         force=True,
         frame_paths=frame_paths,
@@ -79,8 +146,8 @@ def rename_dockerfile(path: Path, new_name: str):
 
 if __name__ == "__main__":
     client = FutureHouseClient(
-        # stage=Stage.from_string(os.environ.get("CROW_ENV", ENV_VARS["STAGE"])),
-        stage=Stage.from_string(os.environ.get("CROW_ENV", "LOCAL")),
+        stage=Stage.from_string(os.environ.get("CROW_ENV", ENV_VARS["STAGE"])),
+        # stage=Stage.from_string(os.environ.get("CROW_ENV", "LOCAL")),
         organization="FutureHouse",
         auth_type=AuthType.API_KEY,
         api_key=os.environ[f"CROW_API_KEY_{ENV_VARS['STAGE']}"],