Update for initial release

.github/workflows/propose_release.yml

@@ -14,15 +14,15 @@ jobs:
     uses: neongeckocom/.github/.github/workflows/propose_semver_release.yml@master
     with:
       branch: dev
-      release_type: ${ inputs.release_type }
+      release_type: ${{ inputs.release_type }}
       update_changelog: True
-      version_file: "neon_inference_plugin_rag/version.py"
+      version_file: "brainforge_inference_plugin_rag/version.py"
   pull_changes:
     uses: neongeckocom/.github/.github/workflows/pull_master.yml@master
     needs: update_version
     with:
       pr_reviewer: neonreviewers
-      pr_assignee: ${ github.actor }
+      pr_assignee: ${{ github.actor }}
       pr_draft: false
-      pr_title: ${ needs.update_version.outputs.version }
-      pr_body: ${ needs.update_version.outputs.changelog }
+      pr_title: ${{ needs.update_version.outputs.version }}
+      pr_body: ${{ needs.update_version.outputs.changelog }}

.github/workflows/publish_release.yml

@@ -1,19 +1,12 @@
-name: Publish GitHub Release
+# This workflow will generate a release distribution and upload it to PyPI
+
+name: Publish Build and GitHub Release
 on:
   push:
     branches:
       - master
 
 jobs:
-  tag_release:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-      - name: Get Version
-        run: |
-          VERSION=$(python setup.py --version)
-          echo "VERSION=${VERSION}" >> $GITHUB_ENV
-      - uses: ncipollo/release-action@v1
-        with:
-          token: ${{secrets.GITHUB_TOKEN}}
-          tag: ${{env.VERSION}}
+  build_and_publish_pypi_and_release:
+    uses: neongeckocom/.github/.github/workflows/publish_stable_release.yml@master
+    secrets: inherit

.github/workflows/publish_test_build.yml

@@ -6,13 +6,13 @@ on:
     branches:
       - dev
     paths-ignore:
-      - 'neon_inference_plugin_rag/version.py'
+      - 'brainforge_inference_plugin_rag/version.py'
 
 jobs:
   publish_alpha_release:
     uses: neongeckocom/.github/.github/workflows/publish_alpha_release.yml@master
     secrets: inherit
     with:
-      publish_pypi: false
-      version_file: "neon_inference_plugin_rag/version.py"
+      publish_pypi: true
+      version_file: "brainforge_inference_plugin_rag/version.py"
       publish_prerelease: true

README.md

@@ -1,24 +1,24 @@
-# Neon Inference Plugin: RAG
+# BrainForge Inference Plugin: RAG
 
 Request step that retrieves documents for selected RAG providers, and a
 response step that rewrites the generated answer with citations. Both steps
-ship together because citations read `request_step_context["neon.rag"]` and
+ship together because citations read `request_step_context["brainforge.rag"]` and
 share `apply_citation_markup`.
 
 | Pipeline | Entry point group | Step id | Class |
 | --- | --- | --- | --- |
-| Request | `brainforge.inference.request_steps` | `neon.rag` | `RagStep` |
-| Response | `brainforge.inference.response_steps` | `neon.citations` | `ResponseParser` |
+| Request | `brainforge.inference.request_steps` | `brainforge.rag` | `RagStep` |
+| Response | `brainforge.inference.response_steps` | `brainforge.citations` | `ResponseParser` |
 
 ```yaml
 pipeline:
   request:
-    - id: neon.rag
+    - id: brainforge.rag
       config:
         mcp_url: https://mcp.neonaiservices2.com/mcp
         top_k: 10
   response:
-    - id: neon.citations
+    - id: brainforge.citations
       config:
         mcp_url: https://mcp.neonaiservices2.com/mcp
         openai_url: https://hana.example/brainforge/openai/chat/completions

neon_inference_plugin_rag/__init__.py → brainforge_inference_plugin_rag/__init__.py

@@ -15,12 +15,12 @@
 
 """RAG retrieval and citation steps for BrainForge inference pipelines."""
 
-from neon_inference_plugin_rag.rag import (
+from brainforge_inference_plugin_rag.rag import (
     RagStep,
     ResponseParser,
     apply_citation_markup,
 )
-from neon_inference_plugin_rag.version import __version__
+from brainforge_inference_plugin_rag.version import __version__
 
 __all__ = [
     "RagStep",

neon_inference_plugin_rag/rag.py → brainforge_inference_plugin_rag/rag.py

@@ -16,7 +16,7 @@
 RAG retrieval and citation rewrite.
 
 `RagStep` lists providers and injects retrieved documents on the request
-pipeline. `ResponseParser` (id `neon.citations`) runs on the response pipeline:
+pipeline. `ResponseParser` (id `brainforge.citations`) runs on the response pipeline:
 it filters those documents, builds a citation completion via MCP, POSTs it
 to a configured OpenAI-compatible endpoint, and writes the annotated text
 back onto the same `UserTransaction`.
@@ -30,7 +30,6 @@
 from typing import Any, Dict, List, Optional, Tuple
 
 import requests
-from neon_data_models.models.api.llm import LLMResponse
 
 from brainforge_inference_utils.mcp import mcp_client, mcp_tool_content_parts
 from brainforge_inference_utils.steps.base import (
@@ -59,7 +58,7 @@ class RagStep(RequestStep):
     - `rag_top_k`: Optional per-request override of configured `top_k`.
     """
 
-    id = "neon.rag"
+    id = "brainforge.rag"
 
     def __init__(self, config: Optional[dict] = None) -> None:
         resolved = dict(config or {})
@@ -252,7 +251,7 @@ class ResponseParser(ResponseStep):
     - `citations_enabled`: When false, the step is skipped.
     """
 
-    id = "neon.citations"
+    id = "brainforge.citations"
     request_context_options = (
         RequestContextOption(
             "citations_enabled",
@@ -282,9 +281,7 @@ def applicable(self, transaction: UserTransaction) -> Optional[str]:
             return "no rag documents"
         return None
 
-    async def process(
-        self, transaction: UserTransaction
-    ) -> Optional[LLMResponse]:
+    async def process(self, transaction: UserTransaction) -> None:
         documents = list(
             (transaction.request_step_context.get(RagStep.id) or {}).get("documents") or [])
         original = transaction.response.response
@@ -297,7 +294,7 @@ async def process(
                     "credits": "",
                     "formatted_credits": "",
                 }
-                return None
+                return
             payload = await self._build_citation_request(
                 original, filtered, timeout)
             raw_cited = await self._complete(
@@ -310,7 +307,7 @@ async def process(
                 "documents": documents,
                 "error": str(exc),
             }
-            return None
+            return
 
         transaction.response_step_context[self.id] = {
             "documents": filtered,
@@ -321,8 +318,8 @@ async def process(
         history = list(transaction.response.history)
         if history and history[-1][0] == "llm":
             history[-1] = ("llm", rewritten)
-        return transaction.response.model_copy(
-            update={"response": rewritten, "history": history})
+        transaction.response.response = rewritten
+        transaction.response.history = history
 
     async def _filter_sources(
         self, response: str, documents: List[dict], timeout: float

requirements/requirements.txt

@@ -1,2 +1,2 @@
-brainforge-inference-utils @ git+https://github.com/NeonGeckoCom/brainforge-inference-utils
+brainforge-inference-utils~=0.0,>=0.0.1a1
 requests

setup.py

@@ -52,7 +52,7 @@ def get_requirements(requirements_filename: str):
     long_description = f.read()
 
 with open(
-    path.join(BASE_PATH, "neon_inference_plugin_rag", "version.py"),
+    path.join(BASE_PATH, "brainforge_inference_plugin_rag", "version.py"),
     "r",
     encoding="utf-8",
 ) as v:
@@ -65,12 +65,12 @@ def get_requirements(requirements_filename: str):
 
 
 setup(
-    name="neon-inference-plugin-rag",
+    name="brainforge-inference-plugin-rag",
     version=version,
     description="RAG retrieval and citation steps for BrainForge inference pipelines",
     long_description=long_description,
     long_description_content_type="text/markdown",
-    url="https://github.com/NeonGeckoCom/neon-inference-plugin-rag",
+    url="https://github.com/NeonGeckoCom/brainforge-inference-plugin-rag",
     author="NeonGecko",
     author_email="developers@neon.ai",
     license="AGPL-3.0-only",
@@ -87,10 +87,10 @@ def get_requirements(requirements_filename: str):
     ],
     entry_points={
         "brainforge.inference.request_steps": [
-            "neon.rag = neon_inference_plugin_rag:RagStep",
+            "brainforge.rag = brainforge_inference_plugin_rag:RagStep",
         ],
         "brainforge.inference.response_steps": [
-            "neon.citations = neon_inference_plugin_rag:ResponseParser",
+            "brainforge.citations = brainforge_inference_plugin_rag:ResponseParser",
         ],
     },
 )

tests/test_entry_point.py

@@ -15,26 +15,26 @@
 
 from importlib.metadata import entry_points
 
-from neon_inference_plugin_rag import RagStep
-from neon_inference_plugin_rag import ResponseParser
+from brainforge_inference_plugin_rag import RagStep
+from brainforge_inference_plugin_rag import ResponseParser
 
 def test_entry_point_loads_rag_step():
     matches = [
         ep for ep in entry_points(group="brainforge.inference.request_steps")
-        if ep.name == "neon.rag"
+        if ep.name == "brainforge.rag"
     ]
     assert len(matches) == 1
     loaded = matches[0].load()
     assert loaded is RagStep
-    assert loaded.id == "neon.rag"
+    assert loaded.id == "brainforge.rag"
 
 
 def test_entry_point_loads_citation_step():
     matches = [
         ep for ep in entry_points(group="brainforge.inference.response_steps")
-        if ep.name == "neon.citations"
+        if ep.name == "brainforge.citations"
     ]
     assert len(matches) == 1
     loaded = matches[0].load()
     assert loaded is ResponseParser
-    assert loaded.id == "neon.citations"
+    assert loaded.id == "brainforge.citations"

tests/test_rag.py

@@ -16,7 +16,7 @@
 import unittest
 from unittest.mock import AsyncMock, patch
 
-from neon_inference_plugin_rag import RagStep
+from brainforge_inference_plugin_rag import RagStep
 from brainforge_inference_utils.transaction import UserTransaction
 from helpers import make_request
 
@@ -59,9 +59,9 @@ async def test_injects_information_and_stores_documents(self):
         self.assertIn("https://b: chunk two", transaction.request.history[0][1])
         self.assertEqual(transaction.request.history[1], ("user", "hi"))
         self.assertEqual(
-            transaction.request_step_context["neon.rag"]["documents"], DOCUMENTS)
+            transaction.request_step_context["brainforge.rag"]["documents"], DOCUMENTS)
         self.assertEqual(
-            transaction.request_step_context["neon.rag"]["providers"], ["news"])
+            transaction.request_step_context["brainforge.rag"]["providers"], ["news"])
 
     async def test_uses_request_top_k_override(self):
         step = RagStep(STEP_CONFIG)
@@ -78,7 +78,7 @@ async def test_uses_request_top_k_override(self):
         lookup.assert_awaited_once_with(
             transaction.request.query, ["docs"], 3, STEP_CONFIG["timeout"])
         self.assertEqual(transaction.request.history, make_request().history)
-        self.assertEqual(transaction.request_step_context["neon.rag"]["documents"], [])
+        self.assertEqual(transaction.request_step_context["brainforge.rag"]["documents"], [])
 
     async def test_lookup_failure_leaves_history_unchanged(self):
         step = RagStep(STEP_CONFIG)
@@ -97,7 +97,7 @@ async def test_lookup_failure_leaves_history_unchanged(self):
 
         self.assertEqual(transaction.request.history, original)
         self.assertEqual(
-            transaction.request_step_context["neon.rag"]["error"], "timed out")
+            transaction.request_step_context["brainforge.rag"]["error"], "timed out")
 
     def test_skips_when_no_providers_selected(self):
         step = RagStep(STEP_CONFIG)

tests/test_response_parser.py

@@ -16,7 +16,7 @@
 import unittest
 from unittest.mock import AsyncMock, patch
 
-from neon_inference_plugin_rag import (
+from brainforge_inference_plugin_rag import (
     ResponseParser,
     apply_citation_markup,
 )
@@ -56,7 +56,7 @@ def _transaction(self, text="See the docs.", **request_context):
             request=request,
             response=make_response(text, request),
             request_context=request_context,
-            request_step_context={"neon.rag": {"documents": DOCUMENTS}},
+            request_step_context={"brainforge.rag": {"documents": DOCUMENTS}},
         )
         return transaction
 
@@ -73,16 +73,16 @@ async def test_rewrites_response_via_openai_endpoint(self):
             step, "_complete",
             new=AsyncMock(return_value="See [1] and [3]."),
         ) as complete:
-            revised = await step.process(transaction)
+            await step.process(transaction)
 
         build.assert_awaited_once()
         complete.assert_awaited_once_with(
             {"messages": []}, transaction.request.model, 30.0)
-        self.assertIsNotNone(revised)
-        self.assertNotEqual(revised.response, "See the docs.")
-        self.assertEqual(revised.history[-1][0], "llm")
-        self.assertEqual(revised.history[-1][1], revised.response)
-        self.assertIn("Doc A", transaction.response_step_context["neon.citations"]["credits"])
+        self.assertNotEqual(transaction.response.response, "See the docs.")
+        self.assertEqual(transaction.response.history[-1][0], "llm")
+        self.assertEqual(
+            transaction.response.history[-1][1], transaction.response.response)
+        self.assertIn("Doc A", transaction.response_step_context["brainforge.citations"]["credits"])
 
     async def test_skips_without_rag_documents(self):
         step = ResponseParser(STEP_CONFIG)
@@ -116,12 +116,11 @@ async def test_completion_failure_leaves_response(self):
             step, "_complete",
             new=AsyncMock(side_effect=RuntimeError("nope")),
         ):
-            revised = await step.process(transaction)
+            await step.process(transaction)
 
-        self.assertIsNone(revised)
         self.assertEqual(transaction.response.response, original)
         self.assertEqual(
-            transaction.response_step_context["neon.citations"]["error"], "nope")
+            transaction.response_step_context["brainforge.citations"]["error"], "nope")
 
     def test_appends_chat_completions_to_base_url(self):
         step = ResponseParser({