Trajectory nodes (#30)

* Set parent and children on transition

* Don't convert to X | Y for type annotations

* Rollback to use Optional as X | Y breaks compability to external libs

* Support save and read trajectories in a tree structure

* Verify save/load trajectories

* Refactor

.env.example

@@ -1,4 +1,9 @@
-INDEX_STORE_DIR=/tmp/moatless/index-store
+MOATLESS_DIR=/tmp/moatless
+
 REPO_DIR=/tmp/repos
-TRAJECTORY_DIR=/tmp/moatless/trajectories
-PRONPT_LOG_DIR=/tmp/moatless/prompt_logs
+
+DEFAULT_MODEL=gpt-4o-2024-05-13
+CHEAP_MODEL=gpt-4o-mini-2024-07-18
+
+INDEX_STORE_DIR=/tmp/moatless/index-store
+INDEX_STORE_URL="https://stmoatless.blob.core.windows.net/indexstore/20240522-voyage-code-2"

.github/workflows/run-tests.yml

@@ -10,14 +10,35 @@ jobs:
   test:
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
     - name: Set up Python
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: '3.10'
     - name: Install Poetry
       uses: snok/install-poetry@v1
     - name: Install dependencies
       run: poetry install
+
+    - name: Cache REPO_DIR
+      uses: actions/cache@v4
+      with:
+        path: /tmp/cache
+        key: ${{ runner.os }}-moatless-${{ github.sha }}
+
     - name: Run tests
-      run: poetry run pytest
+      env:
+        OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        VOYAGE_API_KEY: ${{ secrets.VOYAGE_API_KEY }}
+        INDEX_STORE_URL: ${{ vars.INDEX_STORE_URL }}
+        INDEX_STORE_DIR: /tmp/cache/index-store
+        REPO_DIR: /tmp/cache/repos
+        MOATLESS_DIR: /tmp/moatless
+      run: poetry run pytest
+
+    - name: Upload MOATLESS_DIR contents
+      uses: actions/upload-artifact@v4
+      if: ${{ success() && hashFiles('/tmp/moatless/**') != '' }}
+      with:
+        name: moatless-artifacts
+        path: /tmp/moatless

moatless/__init__.py

@@ -1,3 +1,4 @@
-from moatless.loop import AgenticLoop, Transitions
 from moatless.repository import FileRepository
 from moatless.workspace import Workspace
+from moatless.transition_rules import TransitionRules
+from moatless.loop import AgenticLoop

moatless/benchmark/claude_evaluation.py

@@ -10,8 +10,8 @@
 from moatless.edit.plan import PlanToCode
 from moatless.find.decide import DecideRelevance
 from moatless.find.identify import IdentifyCode
-from moatless.find.search_v2 import SearchCode
-from moatless.loop import Transition
+from moatless.find.search import SearchCode
+from moatless.transition_rules import TransitionRule
 from moatless.state import Finished, Rejected
 from moatless.transitions import (
     search_and_code_transitions,
@@ -177,8 +177,8 @@ def evaluate_search():
         },
         initial_state=SearchCode,
         transitions=[
-            Transition(source=SearchCode, dest=Finished, trigger="did_search"),
-            Transition(source=SearchCode, dest=Finished, trigger="finish"),
+            TransitionRule(source=SearchCode, dest=Finished, trigger="did_search"),
+            TransitionRule(source=SearchCode, dest=Finished, trigger="finish"),
         ],
     )
 
@@ -298,19 +298,19 @@ def evaluate_plan(previous_trajectory_dir: Optional[str] = None):
         },
         initial_state=SearchCode,
         transitions=[
-            Transition(source=SearchCode, dest=IdentifyCode, trigger="did_search"),
-            Transition(source=IdentifyCode, dest=SearchCode, trigger="search"),
-            Transition(source=IdentifyCode, dest=DecideRelevance, trigger="finish"),
-            Transition(source=DecideRelevance, dest=SearchCode, trigger="search"),
-            Transition(
+            TransitionRule(source=SearchCode, dest=IdentifyCode, trigger="did_search"),
+            TransitionRule(source=IdentifyCode, dest=SearchCode, trigger="search"),
+            TransitionRule(source=IdentifyCode, dest=DecideRelevance, trigger="finish"),
+            TransitionRule(source=DecideRelevance, dest=SearchCode, trigger="search"),
+            TransitionRule(
                 source=DecideRelevance,
                 dest=PlanToCode,
                 trigger="finish",
                 exclude_fields={"message"},
             ),
-            Transition(source=PlanToCode, dest=Finished, trigger="edit_code"),
-            Transition(source=PlanToCode, dest=Rejected, trigger="finish"),
-            Transition(source=PlanToCode, dest=Rejected, trigger="reject"),
+            TransitionRule(source=PlanToCode, dest=Finished, trigger="edit_code"),
+            TransitionRule(source=PlanToCode, dest=Rejected, trigger="finish"),
+            TransitionRule(source=PlanToCode, dest=Rejected, trigger="reject"),
         ],
     )