Merge branch 'main' into chore/eval-docs

.github/workflows/prepare_release.yml

@@ -61,4 +61,4 @@ jobs:
           gh pr create -B main --title "$COMMIT_MESSAGE" \
           --body 'Update ${{ github.event.inputs.packageName }} version from ${{ steps.packages_update.outputs.old_version }} to ${{ steps.packages_update.outputs.new_version }}'
         env:
-          GH_TOKEN: ${{ secrets.GH_TOKEN }}
+          GH_TOKEN: ${{ secrets.GH_TOKEN }}

.github/workflows/push_release.yml

@@ -48,4 +48,4 @@ jobs:
           uv tool run twine upload dist/*
         env:
             TWINE_USERNAME: __token__
-            TWINE_PASSWORD: ${{ secrets.PYPI_TOKEN }}
+            TWINE_PASSWORD: ${{ secrets.PYPI_TOKEN }}

docs/how-to/document_search/distributed_ingestion.md

@@ -49,7 +49,7 @@ job_id = client.submit_job(
     runtime_env={
         "working_dir": "./",
         "pip": [
-            "ragbits-core[litellm]",
+            "ragbits-core",
             "ragbits-document-search[distributed]"
         ]
     },
@@ -62,7 +62,7 @@ Ray Jobs is also available as CLI commands. You can submit a job using the follo
 ```bash
 ray job submit \
     --address http://<cluster_address>:8265 \
-    --runtime-env '{"pip": ["ragbits-core[litellm]", "ragbits-document-search[distributed]"]}'\
+    --runtime-env '{"pip": ["ragbits-core", "ragbits-document-search[distributed]"]}'\
     --working-dir . \
     -- python script.py
 ```

docs/quickstart/quickstart1_prompts.md

@@ -7,10 +7,10 @@ In this Quickstart guide, you will learn how to define a dynamic prompt in Ragbi
 To install Ragbits, run the following command in your terminal:
 
 ```bash
-pip install ragbits[litellm]
+pip install ragbits
 ```
 
-This command will install all the popular Ragbits packages, along with [LiteLLM](https://docs.litellm.ai/docs/), which we will use in this guide for communicating with LLM APIs.
+This command will install all the popular Ragbits packages.
 
 ## Defining a Static Prompt
 The most standard way to define a prompt in Ragbits is to create a class that inherits from the `Prompt` class and configure it by setting values for appropriate properties. Here is an example of a simple prompt that asks the model to write a song about Ragbits:

examples/apps/documents_chat.py

@@ -3,7 +3,7 @@
 # dependencies = [
 #     "gradio",
 #     "ragbits-document-search",
-#     "ragbits-core[chroma,litellm]",
+#     "ragbits-core[chroma]",
 # ]
 # ///
 from collections.abc import AsyncIterator

examples/core/llm.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.10"
 # dependencies = [
-#     "ragbits-core[litellm]",
+#     "ragbits-core",
 # ]
 # ///
 import asyncio

examples/document-search/basic.py

@@ -26,7 +26,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[litellm]",
+#     "ragbits-core",
 # ]
 # ///
 

examples/document-search/chroma.py

@@ -27,7 +27,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[chroma,litellm]",
+#     "ragbits-core[chroma]",
 # ]
 # ///
 

examples/document-search/chroma_otel.py

@@ -45,7 +45,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[chroma,litellm,otel]",
+#     "ragbits-core[chroma,otel]",
 # ]
 # ///
 

examples/document-search/distributed.py

@@ -25,7 +25,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search[distributed]",
-#     "ragbits-core[litellm]",
+#     "ragbits-core",
 # ]
 # ///
 

examples/document-search/from_config.py

@@ -24,7 +24,7 @@ class to rephrase the query.
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[chroma,litellm]",
+#     "ragbits-core[chroma]",
 # ]
 # ///
 

examples/document-search/multimodal.py

@@ -27,7 +27,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[litellm]",
+#     "ragbits-core",
 # ]
 # ///
 import asyncio

examples/document-search/qdrant.py

@@ -27,7 +27,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search",
-#     "ragbits-core[litellm,qdrant]",
+#     "ragbits-core[qdrant]",
 # ]
 # ///
 

examples/evaluation/document-search/evaluate.py

@@ -3,7 +3,7 @@
 # dependencies = [
 #     "ragbits-document-search",
 #     "ragbits-evaluate[relari]",
-#     "ragbits-core[litellm,chroma]",
+#     "ragbits-core[chroma]",
 # ]
 # ///
 import asyncio

examples/evaluation/document-search/ingest.py

@@ -2,7 +2,7 @@
 # requires-python = ">=3.10"
 # dependencies = [
 #     "ragbits-document-search[huggingface]",
-#     "ragbits-core[litellm,chroma]",
+#     "ragbits-core[chroma]",
 #     "hydra-core~=1.3.2",
 #     "unstructured[md]>=0.15.13",
 # ]

packages/ragbits-cli/src/ragbits/cli/__init__.py

@@ -1,12 +1,29 @@
 import importlib.util
 import pkgutil
 from pathlib import Path
+from typing import Annotated
 
-from typer import Typer
+import typer
 
 import ragbits
 
-app = Typer(no_args_is_help=True)
+from .app import CLI, OutputType
+
+app = CLI(no_args_is_help=True)
+
+
+@app.callback()
+def output_type(
+    # `OutputType.text.value` used as a workaround for the issue with `typer.Option` not accepting Enum values
+    output: Annotated[
+        OutputType, typer.Option("--output", "-o", help="Set the output type (text or json)")
+    ] = OutputType.text.value,  # type: ignore
+) -> None:
+    """Sets an output type for the CLI
+    Args:
+        output: type of output to be set
+    """
+    app.set_output_type(output_type=output)
 
 
 def main() -> None:

packages/ragbits-cli/src/ragbits/cli/app.py

@@ -0,0 +1,76 @@
+import json
+from dataclasses import dataclass
+from enum import Enum
+from typing import Any
+
+import typer
+from pydantic import BaseModel
+from rich.console import Console
+from rich.table import Table
+
+
+class OutputType(Enum):
+    """Indicates a type of CLI output formatting"""
+
+    text = "text"
+    json = "json"
+
+
+@dataclass()
+class CliState:
+    """A dataclass describing CLI state"""
+
+    output_type: OutputType = OutputType.text
+
+
+class CLI(typer.Typer):
+    """A CLI class with output formatting"""
+
+    def __init__(self, *args: Any, **kwargs: Any):  # noqa: ANN401
+        super().__init__(*args, **kwargs)
+        self.state: CliState = CliState()
+        self.console: Console = Console()
+
+    def set_output_type(self, output_type: OutputType) -> None:
+        """
+        Set the output type in the app state
+        Args:
+            output_type: OutputType
+        """
+        self.state.output_type = output_type
+
+    def print_output(self, data: list[BaseModel] | BaseModel) -> None:
+        """
+        Process and display output based on the current state's output type.
+
+        Args:
+            data: list of ditionaries or list of pydantic models representing output of CLI function
+        """
+        if isinstance(data, BaseModel):
+            data = [data]
+        if len(data) == 0:
+            self._print_empty_list()
+            return
+        first_el_instance = type(data[0])
+        if any(not isinstance(datapoint, first_el_instance) for datapoint in data):
+            raise ValueError("All the rows need to be of the same type")
+        data_dicts: list[dict] = [output.model_dump(mode="python") for output in data]
+        output_type = self.state.output_type
+        if output_type == OutputType.json:
+            print(json.dumps(data_dicts, indent=4))
+        elif output_type == OutputType.text:
+            table = Table(show_header=True, header_style="bold magenta")
+            properties = data[0].model_json_schema()["properties"]
+            for key in properties:
+                table.add_column(properties[key]["title"])
+            for row in data_dicts:
+                table.add_row(*[str(value) for value in row.values()])
+            self.console.print(table)
+        else:
+            raise ValueError(f"Output type: {output_type} not supported")
+
+    def _print_empty_list(self) -> None:
+        if self.state.output_type == OutputType.text:
+            print("Empty data list")
+        elif self.state.output_type == OutputType.json:
+            print(json.dumps([]))

packages/ragbits-core/pyproject.toml

@@ -36,6 +36,7 @@ dependencies = [
     "pydantic>=2.9.1",
     "typer~=0.12.5",
     "tomli~=2.0.2",
+    "litellm~=1.46.0",
 ]
 
 [project.urls]
@@ -48,9 +49,6 @@ dependencies = [
 chroma = [
     "chromadb~=0.4.24",
 ]
-litellm = [
-    "litellm~=1.46.0",
-]
 local = [
     "torch~=2.2.1",
     "transformers~=4.44.2",

packages/ragbits-core/src/ragbits/core/cli.py

@@ -6,11 +6,12 @@
 from pathlib import Path
 
 import typer
-from rich import print as pprint
+from pydantic import BaseModel
 
+from ragbits.cli.app import CLI
 from ragbits.core.config import core_config
 from ragbits.core.llms.base import LLMType
-from ragbits.core.prompt.prompt import Prompt
+from ragbits.core.prompt.prompt import ChatFormat, Prompt
 
 
 def _render(prompt_path: str, payload: str | None) -> Prompt:
@@ -25,10 +26,17 @@ def _render(prompt_path: str, payload: str | None) -> Prompt:
     return prompt_cls()
 
 
+class LLMResponseCliOutput(BaseModel):
+    """An output model for llm responses in CLI"""
+
+    question: ChatFormat
+    answer: str | BaseModel | None = None
+
+
 prompts_app = typer.Typer(no_args_is_help=True)
 
 
-def register(app: typer.Typer) -> None:
+def register(app: CLI) -> None:
     """
     Register the CLI commands for the package.
 
@@ -68,9 +76,8 @@ def render(prompt_path: str, payload: str | None = None) -> None:
         Renders a prompt by loading a class from a module and initializing it with a given payload.
         """
         prompt = _render(prompt_path=prompt_path, payload=payload)
-
-        pprint("[orange3]RENDERED PROMPT:")
-        pprint(prompt.chat)
+        response = LLMResponseCliOutput(question=prompt.chat)
+        app.print_output(response)
 
     @prompts_app.command(name="exec")
     def execute(
@@ -92,11 +99,8 @@ def execute(
             raise ValueError("`llm_factory` must be provided")
         llm = get_llm_from_factory(llm_factory)
 
-        response = asyncio.run(llm.generate(prompt))
-
-        pprint("[orange3]QUESTION:")
-        pprint(prompt.chat)
-        pprint("[orange3]ANSWER:")
-        pprint(response)
+        llm_output = asyncio.run(llm.generate(prompt))
+        response = LLMResponseCliOutput(question=prompt.chat, answer=llm_output)
+        app.print_output(response)
 
     app.add_typer(prompts_app, name="prompts", help="Commands for managing prompts")

packages/ragbits-core/src/ragbits/core/embeddings/litellm.py

@@ -1,9 +1,4 @@
-try:
-    import litellm
-
-    HAS_LITELLM = True
-except ImportError:
-    HAS_LITELLM = False
+import litellm
 
 from ragbits.core.audit import trace
 from ragbits.core.embeddings import Embeddings
@@ -40,13 +35,7 @@ def __init__(
                 for more information, follow the instructions for your specific vendor in the\
                 [LiteLLM documentation](https://docs.litellm.ai/docs/embedding/supported_embedding).
             api_version: The API version for the call.
-
-        Raises:
-            ImportError: If the 'litellm' extra requirements are not installed.
         """
-        if not HAS_LITELLM:
-            raise ImportError("You need to install the 'litellm' extra requirements to use LiteLLM embeddings models")
-
         super().__init__()
         self.model = model
         self.options = options or {}

packages/ragbits-core/src/ragbits/core/llms/__init__.py

@@ -3,8 +3,9 @@
 from ragbits.core.utils.config_handling import get_cls_from_config
 
 from .base import LLM
+from .litellm import LiteLLM
 
-__all__ = ["LLM"]
+__all__ = ["LLM", "LiteLLM"]
 
 module = sys.modules[__name__]