Merge pull request #1043 from Pythagora-io/crashes

Crashes

core/agents/code_monkey.py

@@ -57,10 +57,12 @@ async def implement_changes(self) -> AgentResponse:
         user_feedback = None
         user_feedback_qa = None
         llm = self.get_llm()
+
         if iterations:
-            instructions = iterations[-1]["description"]
-            user_feedback = iterations[-1]["user_feedback"]
-            user_feedback_qa = iterations[-1]["user_feedback_qa"]
+            last_iteration = iterations[-1]
+            instructions = last_iteration.get("description")
+            user_feedback = last_iteration.get("user_feedback")
+            user_feedback_qa = last_iteration.get("user_feedback_qa")
         else:
             instructions = self.current_state.current_task["instructions"]
 

core/agents/developer.py

@@ -181,12 +181,11 @@ async def breakdown_current_task(self) -> AgentResponse:
         )
         response: str = await llm(convo)
 
-        # FIXME: check if this is correct, as sqlalchemy can't figure out modifications
-        # to attributes; however, self.next is not saved yet so maybe this is fine
         self.next_state.tasks[current_task_index] = {
             **task,
             "instructions": response,
         }
+        self.next_state.flag_tasks_as_modified()
 
         await self.send_message("Breaking down the task into steps ...")
         convo.assistant(response).template("parse_task").require_schema(TaskSteps)

core/agents/tech_lead.py

@@ -17,19 +17,19 @@
 log = get_logger(__name__)
 
 
-class Task(BaseModel):
-    description: str = Field(description=("Very detailed description of a development task."))
+class Epic(BaseModel):
+    description: str = Field(description=("Description of an epic."))
 
 
 class DevelopmentPlan(BaseModel):
-    plan: list[Task] = Field(description="List of development tasks that need to be done to implement the entire plan.")
+    plan: list[Epic] = Field(description="List of epics that need to be done to implement the entire plan.")
 
 
 class UpdatedDevelopmentPlan(BaseModel):
-    updated_current_epic: Task = Field(
-        description="Updated detailed description of what was implemented while working on the current development task."
+    updated_current_epic: Epic = Field(
+        description="Updated description of what was implemented while working on the current epic."
     )
-    plan: list[Task] = Field(description="List of unfinished development tasks.")
+    plan: list[Epic] = Field(description="List of unfinished epics.")
 
 
 class TechLead(BaseAgent):

core/agents/troubleshooter.py

@@ -159,11 +159,7 @@ async def _get_route_files(self) -> list[File]:
         """Returns the list of file paths that have routes defined in them."""
 
         llm = self.get_llm(ROUTE_FILES_AGENT_NAME)
-        convo = (
-            self._get_task_convo()
-            .template("get_route_files", task=self.current_state.current_task)
-            .require_schema(RouteFilePaths)
-        )
+        convo = AgentConvo(self).template("get_route_files").require_schema(RouteFilePaths)
         file_list = await llm(convo, parser=JSONParser(RouteFilePaths))
         route_files: set[str] = set(file_list.files)
 

core/prompts/architect/technologies.prompt

@@ -1,6 +1,6 @@
 You're designing the architecture and technical specifications for a new project.
 
-If the project requirements call out for specific technology, use that. Otherwise, if working on a web app, prefer Node.js for the backend (with Express if a web server is needed, and MongoDB if a database is needed), and Bootstrap for the front-end. You MUST NOT use Docker, Kubernetes, microservices and single-page app frameworks like React, Next.js, Angular, Vue or Svelte unless the project details explicitly require it.
+If the project requirements call out for specific technology, use that. Otherwise, if working on a web app, prefer Node.js for the backend (with Express if a web server is needed, and MongoDB if a database is needed), and Bootstrap for the front-end. You MUST NOT use Docker, Kubernetes, microservices and single-page app frameworks like Next.js, Angular, Vue or Svelte unless the project details explicitly require it.
 
 Here is a high level description of "{{ state.branch.project.name }}":
 ```

core/prompts/tech-lead/system.prompt

@@ -1,4 +1,3 @@
 You are an experienced tech lead in a software development agency.
-Your main task is to break down the project into smaller tasks that developers will do.
-You must specify each task as clear as possible.
-Each task must have a description of what needs to be implemented.
+Your main job is to break down the project into epics that developers will do.
+You must specify each epic as clear as possible.

core/prompts/troubleshooter/get_route_files.prompt

@@ -7,4 +7,7 @@ The solution was to add the entire content of all the files that have routes def
 them, and this prompt selects those files.
 
 #}
+{% include "partials/project_details.prompt" %}
+{% include "partials/files_list.prompt" %}
+
 Your task is to identify all the files, from the above list, that have routes defined in them. Return just the file paths as a JSON list named "files", DO NOT return anything else. If there are no files with routes, return an empty list.

tests/agents/test_tech_lead.py

@@ -1,7 +1,7 @@
 import pytest
 
 from core.agents.response import ResponseType
-from core.agents.tech_lead import DevelopmentPlan, Task, TechLead, UpdatedDevelopmentPlan
+from core.agents.tech_lead import DevelopmentPlan, Epic, TechLead, UpdatedDevelopmentPlan
 from core.db.models import Complexity
 from core.db.models.project_state import TaskStatus
 from core.ui.base import UserInput
@@ -87,8 +87,8 @@ async def test_plan_epic(agentcontext):
     tl.get_llm = mock_get_llm(
         return_value=DevelopmentPlan(
             plan=[
-                Task(description="Task 1"),
-                Task(description="Task 2"),
+                Epic(description="Task 1"),
+                Epic(description="Task 2"),
             ]
         )
     )
@@ -122,8 +122,8 @@ async def test_update_epic(agentcontext):
     tl = TechLead(sm, ui)
     tl.get_llm = mock_get_llm(
         return_value=UpdatedDevelopmentPlan(
-            updated_current_epic=Task(description="Updated Just Finished"),
-            plan=[Task(description="Alternative Future Task")],
+            updated_current_epic=Epic(description="Updated Just Finished"),
+            plan=[Epic(description="Alternative Future Task")],
         )
     )