feat: make spec optional (#719)

* feat: Add spec_enabled configuration for data science settings

* make spec alternative

* change spec logic in exp_gen

* remove some general texts

* align

---------

Co-authored-by: Young <afe.young@gmail.com>
Co-authored-by: yuanteli <1957922024@qq.com>

rdagent/components/coder/data_science/ensemble/__init__.py

@@ -12,8 +12,12 @@ File structure
 """
 
 import json
+from pathlib import Path
 from typing import Dict
 
+from jinja2 import Environment, StrictUndefined
+
+from rdagent.app.data_science.conf import DS_RD_SETTING
 from rdagent.components.coder.CoSTEER import CoSTEER
 from rdagent.components.coder.CoSTEER.evaluators import (
     CoSTEERMultiEvaluator,
@@ -35,6 +39,8 @@ from rdagent.oai.llm_utils import APIBackend
 from rdagent.utils.agent.ret import PythonAgentOut
 from rdagent.utils.agent.tpl import T
 
+DIRNAME = Path(__file__).absolute().resolve().parent
+
 
 class EnsembleMultiProcessEvolvingStrategy(MultiProcessEvolvingStrategy):
     def implement_one_task(
@@ -79,8 +85,24 @@ class EnsembleMultiProcessEvolvingStrategy(MultiProcessEvolvingStrategy):
             all_code=workspace.all_codes,
             out_spec=PythonAgentOut.get_spec(),
         )
+
+        if DS_RD_SETTING.spec_enabled:
+            code_spec = workspace.file_dict["spec/ensemble.md"]
+        else:
+            test_code = (
+                Environment(undefined=StrictUndefined)
+                .from_string((DIRNAME / "eval_tests" / "ensemble_test.txt").read_text())
+                .render(
+                    model_names=[
+                        fn[:-3] for fn in workspace.file_dict.keys() if fn.startswith("model_") and "test" not in fn
+                    ]
+                )
+            )
+            code_spec = T("scenarios.data_science.share:component_spec.general").r(
+                spec=T("scenarios.data_science.share:component_spec.Ensemble").r(), test_code=test_code
+            )
         user_prompt = T(".prompts:ensemble_coder.user").r(
-            ensemble_spec=workspace.file_dict["spec/ensemble.md"],
+            code_spec=code_spec,
             latest_code=workspace.file_dict.get("ensemble.py"),
             latest_code_feedback=prev_task_feedback,
         )

rdagent/components/coder/data_science/ensemble/eval_tests/ensemble_test.txt

@@ -1,9 +1,11 @@
 """
-A qualified ensemble implementation should:
-- Successfully run
+Tests for `ensemble_workflow` in ensemble.py
+
+A qualified ensemble_workflow implementation should:
 - Return predictions
 - Have correct shapes for inputs and outputs
 - Use validation data appropriately
+- Generate a scores.csv file
 """
 
 import numpy as np

rdagent/components/coder/data_science/ensemble/prompts.yaml

@@ -51,8 +51,8 @@ ensemble_coder:
     {% endif %}
 
   user: |-
-    --------- Ensemble Specification ---------
-    {{ ensemble_spec }}
+    --------- Code Specification ---------
+    {{ code_spec }}
 
     {% if latest_code %}
     --------- Former code ---------