API Updates: fleshing out RAG APIs, introduce "llama stack" CLI command (#51)

* add tools to chat completion request * use templates for generating system prompts * Moved ToolPromptFormat and jinja templates to llama_models.llama3.api * <WIP> memory changes - inlined AgenticSystemInstanceConfig so API feels more ergonomic - renamed it to AgentConfig, AgentInstance -> Agent - added a MemoryConfig and `memory` parameter - added `attachments` to input and `output_attachments` to the response - some naming changes * InterleavedTextAttachment -> InterleavedTextMedia, introduce memory tool * flesh out memory banks API * agentic loop has a RAG implementation * faiss provider implementation * memory client works * re-work tool definitions, fix FastAPI issues, fix tool regressions * fix agentic_system utils * basic RAG seems to work * small bug fixes for inline attachments * Refactor custom tool execution utilities * Bug fix, show memory retrieval steps in EventLogger * No need for api_key for Remote providers * add special unicode character ↵ to showcase newlines in model prompt templates * remove api.endpoints imports * combine datatypes.py and endpoints.py into api.py * Attachment / add TTL api * split batch_inference from inference * minor import fixes * use a single impl for ChatFormat.decode_assistant_mesage * use interleaved_text_media_as_str() utilityt * Fix api.datatypes imports * Add blobfile for tiktoken * Add ToolPromptFormat to ChatFormat.encode_message so that tools are encoded properly * templates take optional --format={json,function_tag} * Rag Updates * Add `api build` subcommand -- WIP * fix * build + run image seems to work * <WIP> adapters * bunch more work to make adapters work * api build works for conda now * ollama remote adapter works * Several smaller fixes to make adapters work Also, reorganized the pattern of __init__ inside providers so configuration can stay lightweight * llama distribution -> llama stack + containers (WIP) * All the new CLI for api + stack work * Make Fireworks and Together into the Adapter format * Some quick fixes to the CLI behavior to make it consistent * Updated README phew * Update cli_reference.md * llama_toolchain/distribution -> llama_toolchain/core * Add termcolor * update paths * Add a log just for consistency * chmod +x scripts * Fix api dependencies not getting added to configuration * missing import lol * Delete utils.py; move to agentic system * Support downloading of URLs for attachments for code interpreter * Simplify and generalize `llama api build` yay * Update `llama stack configure` to be very simple also * Fix stack start * Allow building an "adhoc" distribution * Remote `llama api []` subcommands * Fixes to llama stack commands and update docs * Update documentation again and add error messages to llama stack start * llama stack start -> llama stack run * Change name of build for less confusion * Add pyopenapi fork to the repository, update RFC assets * Remove conflicting annotation * Added a "--raw" option for model template printing --------- Co-authored-by: Hardik Shah <hjshah@fb.com> Co-authored-by: Ashwin Bharambe <ashwin@meta.com> Co-authored-by: Dalton Flanagan <6599399+dltn@users.noreply.github.com>
2024-09-03 22:39:39 -07:00 · 2024-09-03 22:39:39 -07:00 · 7bc7785b0d
commit 7bc7785b0d
parent 35093c0b6f
141 changed files with 8252 additions and 4032 deletions
--- a/llama_toolchain/inference/api/api.py
+++ b/llama_toolchain/inference/api/api.py
@ -0,0 +1,187 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from enum import Enum
+
+from typing import List, Literal, Optional, Protocol, Union
+
+from llama_models.schema_utils import json_schema_type, webmethod
+
+from pydantic import BaseModel, Field
+from typing_extensions import Annotated
+
+from llama_models.llama3.api.datatypes import *  # noqa: F403
+
+
+class LogProbConfig(BaseModel):
+    top_k: Optional[int] = 0
+
+
+@json_schema_type
+class QuantizationType(Enum):
+    bf16 = "bf16"
+    fp8 = "fp8"
+
+
+@json_schema_type
+class Fp8QuantizationConfig(BaseModel):
+    type: Literal[QuantizationType.fp8.value] = QuantizationType.fp8.value
+
+
+@json_schema_type
+class Bf16QuantizationConfig(BaseModel):
+    type: Literal[QuantizationType.bf16.value] = QuantizationType.bf16.value
+
+
+QuantizationConfig = Annotated[
+    Union[Bf16QuantizationConfig, Fp8QuantizationConfig],
+    Field(discriminator="type"),
+]
+
+
+@json_schema_type
+class ChatCompletionResponseEventType(Enum):
+    start = "start"
+    complete = "complete"
+    progress = "progress"
+
+
+@json_schema_type
+class ToolCallParseStatus(Enum):
+    started = "started"
+    in_progress = "in_progress"
+    failure = "failure"
+    success = "success"
+
+
+@json_schema_type
+class ToolCallDelta(BaseModel):
+    content: Union[str, ToolCall]
+    parse_status: ToolCallParseStatus
+
+
+@json_schema_type
+class ChatCompletionResponseEvent(BaseModel):
+    """Chat completion response event."""
+
+    event_type: ChatCompletionResponseEventType
+    delta: Union[str, ToolCallDelta]
+    logprobs: Optional[List[TokenLogProbs]] = None
+    stop_reason: Optional[StopReason] = None
+
+
+@json_schema_type
+class CompletionRequest(BaseModel):
+    model: str
+    content: InterleavedTextMedia
+    sampling_params: Optional[SamplingParams] = SamplingParams()
+
+    stream: Optional[bool] = False
+    logprobs: Optional[LogProbConfig] = None
+
+
+@json_schema_type
+class CompletionResponse(BaseModel):
+    completion_message: CompletionMessage
+    logprobs: Optional[List[TokenLogProbs]] = None
+
+
+@json_schema_type
+class CompletionResponseStreamChunk(BaseModel):
+    """streamed completion response."""
+
+    delta: str
+    stop_reason: Optional[StopReason] = None
+    logprobs: Optional[List[TokenLogProbs]] = None
+
+
+@json_schema_type
+class BatchCompletionRequest(BaseModel):
+    model: str
+    content_batch: List[InterleavedTextMedia]
+    sampling_params: Optional[SamplingParams] = SamplingParams()
+    logprobs: Optional[LogProbConfig] = None
+
+
+@json_schema_type
+class BatchCompletionResponse(BaseModel):
+    completion_message_batch: List[CompletionMessage]
+
+
+@json_schema_type
+class ChatCompletionRequest(BaseModel):
+    model: str
+    messages: List[Message]
+    sampling_params: Optional[SamplingParams] = SamplingParams()
+
+    # zero-shot tool definitions as input to the model
+    tools: Optional[List[ToolDefinition]] = Field(default_factory=list)
+    tool_choice: Optional[ToolChoice] = Field(default=ToolChoice.auto)
+    tool_prompt_format: Optional[ToolPromptFormat] = Field(
+        default=ToolPromptFormat.json
+    )
+
+    stream: Optional[bool] = False
+    logprobs: Optional[LogProbConfig] = None
+
+
+@json_schema_type
+class ChatCompletionResponseStreamChunk(BaseModel):
+    """SSE-stream of these events."""
+
+    event: ChatCompletionResponseEvent
+
+
+@json_schema_type
+class ChatCompletionResponse(BaseModel):
+    completion_message: CompletionMessage
+    logprobs: Optional[List[TokenLogProbs]] = None
+
+
+@json_schema_type
+class BatchChatCompletionRequest(BaseModel):
+    model: str
+    messages_batch: List[List[Message]]
+    sampling_params: Optional[SamplingParams] = SamplingParams()
+
+    # zero-shot tool definitions as input to the model
+    tools: Optional[List[ToolDefinition]] = Field(default_factory=list)
+    tool_choice: Optional[ToolChoice] = Field(default=ToolChoice.auto)
+    tool_prompt_format: Optional[ToolPromptFormat] = Field(
+        default=ToolPromptFormat.json
+    )
+    logprobs: Optional[LogProbConfig] = None
+
+
+@json_schema_type
+class BatchChatCompletionResponse(BaseModel):
+    completion_message_batch: List[CompletionMessage]
+
+
+@json_schema_type
+class EmbeddingsResponse(BaseModel):
+    embeddings: List[List[float]]
+
+
+class Inference(Protocol):
+    @webmethod(route="/inference/completion")
+    async def completion(
+        self,
+        request: CompletionRequest,
+    ) -> Union[CompletionResponse, CompletionResponseStreamChunk]: ...
+
+    @webmethod(route="/inference/chat_completion")
+    async def chat_completion(
+        self,
+        request: ChatCompletionRequest,
+    ) -> Union[ChatCompletionResponse, ChatCompletionResponseStreamChunk]: ...
+
+    @webmethod(route="/inference/embeddings")
+    async def embeddings(
+        self,
+        model: str,
+        contents: List[InterleavedTextMedia],
+    ) -> EmbeddingsResponse: ...