feat: add Avian as an LLM provider

.github/workflows/tests.yml

@@ -45,13 +45,13 @@ jobs:
       # Run even if make test fails for coverage reports
       # TODO: select a better xml results displayer
       - name: Archive test results coverage results
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         if: always()
         with:
           name: test_results
           path: tests-results.xml
       - name: Archive code coverage results
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         if: always()
         with:
           name: code-coverage-report

private_gpt/components/embedding/embedding_component.py

@@ -161,6 +161,22 @@ def __init__(self, settings: Settings) -> None:
                     api_key=api_key,
                     model=model,
                 )
+            case "avian":
+                try:
+                    from llama_index.embeddings.openai import (  # type: ignore
+                        OpenAIEmbedding,
+                    )
+                except ImportError as e:
+                    raise ImportError(
+                        "Avian dependencies not found, install with `poetry install --extras embeddings-avian`"
+                    ) from e
+
+                avian_settings = settings.avian
+                self.embedding_model = OpenAIEmbedding(
+                    api_base=avian_settings.api_base,
+                    api_key=avian_settings.api_key,
+                    model=avian_settings.embedding_model,
+                )
             case "mock":
                 # Not a random number, is the dimensionality used by
                 # the default embedding model

private_gpt/components/llm/llm_component.py

@@ -221,5 +221,20 @@ def wrapper(*args: Any, **kwargs: Any) -> Any:
                 self.llm = Gemini(
                     model_name=gemini_settings.model, api_key=gemini_settings.api_key
                 )
+            case "avian":
+                try:
+                    from llama_index.llms.openai import OpenAI  # type: ignore
+                except ImportError as e:
+                    raise ImportError(
+                        "Avian dependencies not found, install with `poetry install --extras llms-avian`"
+                    ) from e
+
+                avian_settings = settings.avian
+                self.llm = OpenAI(
+                    api_base=avian_settings.api_base,
+                    api_key=avian_settings.api_key,
+                    model=avian_settings.model,
+                    timeout=avian_settings.request_timeout,
+                )
             case "mock":
                 self.llm = MockLLM()

private_gpt/settings/settings.py

@@ -115,6 +115,7 @@ class LLMSettings(BaseModel):
         "mock",
         "ollama",
         "gemini",
+        "avian",
     ]
     max_new_tokens: int = Field(
         256,
@@ -205,6 +206,7 @@ class EmbeddingSettings(BaseModel):
         "mock",
         "gemini",
         "mistralai",
+        "avian",
     ]
     ingest_mode: Literal["simple", "batch", "parallel", "pipeline"] = Field(
         "simple",
@@ -280,6 +282,29 @@ class GeminiSettings(BaseModel):
     )
 
 
+class AvianSettings(BaseModel):
+    api_base: str = Field(
+        "https://api.avian.io/v1",
+        description="Base URL of Avian API.",
+    )
+    api_key: str = Field(
+        description="Avian API key. Get one at https://avian.io.",
+    )
+    model: str = Field(
+        "deepseek/deepseek-v3.2",
+        description="Avian model to use. Options: 'deepseek/deepseek-v3.2', "
+        "'moonshotai/kimi-k2.5', 'z-ai/glm-5', 'minimax/minimax-m2.5'.",
+    )
+    embedding_model: str = Field(
+        "text-embedding-ada-002",
+        description="Avian embedding model to use.",
+    )
+    request_timeout: float = Field(
+        120.0,
+        description="Time elapsed until Avian server times out the request. Default is 120s.",
+    )
+
+
 class OllamaSettings(BaseModel):
     api_base: str = Field(
         "http://localhost:11434",
@@ -598,6 +623,7 @@ class Settings(BaseModel):
     sagemaker: SagemakerSettings
     openai: OpenAISettings
     gemini: GeminiSettings
+    avian: AvianSettings
     ollama: OllamaSettings
     azopenai: AzureOpenAISettings
     vectorstore: VectorstoreSettings

pyproject.toml

@@ -69,12 +69,14 @@ llms-ollama = ["llama-index-llms-ollama"]
 llms-sagemaker = ["boto3"]
 llms-azopenai = ["llama-index-llms-azure-openai"]
 llms-gemini = ["llama-index-llms-gemini"]
+llms-avian = ["llama-index-llms-openai"]
 embeddings-ollama = ["llama-index-embeddings-ollama"]
 embeddings-huggingface = ["llama-index-embeddings-huggingface", "einops"]
 embeddings-openai = ["llama-index-embeddings-openai"]
 embeddings-sagemaker = ["boto3"]
 embeddings-azopenai = ["llama-index-embeddings-azure-openai"]
 embeddings-gemini = ["llama-index-embeddings-gemini"]
+embeddings-avian = ["llama-index-embeddings-openai"]
 embeddings-mistral = ["llama-index-embeddings-mistralai"]
 vector-stores-qdrant = ["llama-index-vector-stores-qdrant"]
 vector-stores-clickhouse = ["llama-index-vector-stores-clickhouse", "clickhouse_connect"]

settings-avian.yaml

@@ -0,0 +1,17 @@
+server:
+  env_name: ${APP_ENV:avian}
+
+llm:
+  mode: avian
+  max_new_tokens: 512
+  context_window: 131072
+  temperature: 0.1
+
+embedding:
+  mode: avian
+  ingest_mode: simple
+
+avian:
+  api_base: ${AVIAN_API_BASE:https://api.avian.io/v1}
+  api_key: ${AVIAN_API_KEY:}
+  model: deepseek/deepseek-v3.2

settings.yaml

@@ -150,3 +150,10 @@ gemini:
   api_key: ${GOOGLE_API_KEY:}
   model: models/gemini-pro
   embedding_model: models/embedding-001
+
+avian:
+  api_base: ${AVIAN_API_BASE:https://api.avian.io/v1}
+  api_key: ${AVIAN_API_KEY:}
+  model: deepseek/deepseek-v3.2
+  embedding_model: text-embedding-ada-002
+  request_timeout: 120.0