Adds the ability to pass files to ramalama run

docs/ramalama-chat.1.md

@@ -29,9 +29,13 @@ Show this help message and exit
 #### **--prefix**
 Prefix for the user prompt (default: 🦭 > )
 
+#### **--rag**=path
+A file or directory of files to be loaded and provided as local context in the chat history.
+
 #### **--url**=URL
 The host to send requests to (default: http://127.0.0.1:8080)
 
+
 ## EXAMPLES
 
 Communicate with the default local OpenAI REST API. (http://127.0.0.1:8080)

pyproject.toml

@@ -8,7 +8,6 @@ version = "0.9.3"
 description = "RamaLama is a command line tool for working with AI LLM models."
 readme = "README.md"
 requires-python = ">=3.10"
-license = { file = "LICENSE" }
 keywords = ["ramalama", "llama", "AI"]
 dependencies = [
   "argcomplete",
@@ -18,6 +17,8 @@ maintainers = [
    { name="Eric Curtin", email = "[email protected]" },
 ]
 
+[project.license]
+text = "MIT"
 
 [project.optional-dependencies]
 dev = [
@@ -56,6 +57,7 @@ ramalama = "ramalama.cli:main"
 
 [tool.setuptools]
 include-package-data = true
+license-files = ["LICENSE"]
 
 [tool.black]
 line-length = 120
@@ -93,7 +95,7 @@ log_cli_date_format = "%Y-%m-%d %H:%M:%S"
 
 
 [tool.setuptools.packages.find]
-include = ["ramalama"]
+include = ["ramalama", "ramalama.*"]
 
 [tool.setuptools.data-files]
 "share/ramalama" = ["shortnames/shortnames.conf"]

ramalama/chat.py

@@ -14,6 +14,7 @@
 from ramalama.config import CONFIG
 from ramalama.console import EMOJI, should_colorize
 from ramalama.engine import dry_run, stop_container
+from ramalama.file_upload.file_loader import FileUpLoader
 from ramalama.logger import logger
 
 
@@ -72,6 +73,16 @@ def __init__(self, args):
         self.prompt = args.prefix
 
         self.url = f"{args.url}/chat/completions"
+        self.prep_rag_message()
+
+    def prep_rag_message(self):
+        if (context := getattr(self.args, "rag", None)) is None:
+            return
+
+        if not (message_content := FileUpLoader(context).load()):
+            return
+
+        self.conversation_history.append({"role": "system", "content": message_content})
 
     def handle_args(self):
         prompt = " ".join(self.args.ARGS) if self.args.ARGS else None

ramalama/cli.py

@@ -905,6 +905,7 @@ def chat_parser(subparsers):
         help='possible values are "never", "always" and "auto".',
     )
     parser.add_argument("--prefix", type=str, help="prefix for the user prompt", default=default_prefix())
+    parser.add_argument("--rag", type=str, help="a file or directory to use as context for the chat")
     parser.add_argument("--url", type=str, default="http://127.0.0.1:8080/v1", help="the url to send requests to")
     parser.add_argument("MODEL", completer=local_models)  # positional argument
     parser.add_argument(
@@ -925,6 +926,7 @@ def run_parser(subparsers):
     )
     parser.add_argument("--prefix", type=str, help="prefix for the user prompt", default=default_prefix())
     parser.add_argument("MODEL", completer=local_models)  # positional argument
+
     parser.add_argument(
         "ARGS",
         nargs="*",

ramalama/file_upload/__init__.py

@@ -0,0 +1,3 @@
+from ramalama.file_upload import file_loader, file_types
+
+__all__ = ["file_loader", "file_types"]

ramalama/file_upload/file_loader.py

@@ -0,0 +1,57 @@
+import os
+from string import Template
+from warnings import warn
+
+from ramalama.file_upload.file_types import base, txt
+
+SUPPORTED_EXTENSIONS = {
+    '.txt': txt.TXTFileUpload,
+    '.sh': txt.TXTFileUpload,
+    '.md': txt.TXTFileUpload,
+    '.yaml': txt.TXTFileUpload,
+    '.yml': txt.TXTFileUpload,
+    '.json': txt.TXTFileUpload,
+    '.csv': txt.TXTFileUpload,
+    '.toml': txt.TXTFileUpload,
+}
+
+
+class BaseFileUploader:
+    """
+    Base class for file upload handlers.
+    This class should be extended by specific file type handlers.
+    """
+
+    def __init__(self, files: list[base.BaseFileUpload], delim_string: str = "<!--start_document $name-->"):
+        self.files = files
+        self.document_delimiter: Template = Template(delim_string)
+
+    def load(self) -> str:
+        """
+        Generate the output string by concatenating the processed files.
+        """
+        output = (f"\n{self.document_delimiter.substitute(name=f.file)}\n{f.load()}" for f in self.files)
+        return "".join(output)
+
+
+class FileUpLoader(BaseFileUploader):
+    def __init__(self, file_path: str):
+        if not os.path.exists(file_path):
+            raise ValueError(f"{file_path} does not exist.")
+
+        if not os.path.isdir(file_path):
+            files = [file_path]
+        else:
+            files = [os.path.join(root, name) for root, _, files in os.walk(file_path) for name in files]
+
+        extensions = [os.path.splitext(f)[1].lower() for f in files]
+
+        if set(extensions) - set(SUPPORTED_EXTENSIONS):
+            warning_message = (
+                f"Unsupported file types found: {set(extensions) - set(SUPPORTED_EXTENSIONS)}\n"
+                f"Supported types are: {set(SUPPORTED_EXTENSIONS.keys())}"
+            )
+            warn(warning_message)
+
+        files = [SUPPORTED_EXTENSIONS[ext](file=f) for ext, f in zip(extensions, files) if ext in SUPPORTED_EXTENSIONS]
+        super().__init__(files=files)

ramalama/file_upload/file_types/__init__.py

@@ -0,0 +1,3 @@
+from ramalama.file_upload.file_types import base, txt
+
+__all__ = ["base", "txt"]

ramalama/file_upload/file_types/base.py

@@ -0,0 +1,19 @@
+from abc import ABC, abstractmethod
+
+
+class BaseFileUpload(ABC):
+    """
+    Base class for file upload handlers.
+    This class should be extended by specific file type handlers.
+    """
+
+    def __init__(self, file):
+        self.file = file
+
+    @abstractmethod
+    def load(self) -> str:
+        """
+        Load the content of the file.
+        This method should be implemented by subclasses to handle specific file types.
+        """
+        pass

ramalama/file_upload/file_types/pdf.py

@@ -0,0 +1,10 @@
+from ramalama.file_upload.file_types.base import BaseFileUpload
+
+
+class PDFFileUpload(BaseFileUpload):
+    def load(self) -> str:
+        """
+        Load the content of the PDF file.
+        This method should be implemented to handle PDF file reading.
+        """
+        return ""

ramalama/file_upload/file_types/txt.py

@@ -0,0 +1,12 @@
+from ramalama.file_upload.file_types.base import BaseFileUpload
+
+
+class TXTFileUpload(BaseFileUpload):
+    def load(self) -> str:
+        """
+        Load the content of the text file.
+        """
+
+        # TODO: Support for non-default encodings?
+        with open(self.file, 'r') as f:
+            return f.read()

test/unit/conftest.py

@@ -0,0 +1,6 @@
+# # tests/conftest.py
+# import pytest
+
+# @pytest.fixture(autouse=True)
+# def set_container_engine_env(monkeypatch):
+#     monkeypatch.setenv("RAMALAMA_CONTAINER_ENGINE", "docker")

test/unit/data/test_file_upload/sample.csv

@@ -0,0 +1,6 @@
+name,age,city,occupation
+John Doe,30,New York,Engineer
+Jane Smith,25,San Francisco,Designer
+Bob Johnson,35,Chicago,Manager
+Alice Brown,28,Boston,Developer
+Charlie Wilson,32,Seattle,Analyst 

test/unit/data/test_file_upload/sample.json

@@ -0,0 +1,20 @@
+{
+  "name": "test_data",
+  "version": "1.0.0",
+  "description": "Sample JSON data for testing file upload functionality",
+  "features": [
+    "text_processing",
+    "file_upload",
+    "chat_integration"
+  ],
+  "metadata": {
+    "author": "Test User",
+    "created": "2024-01-01",
+    "tags": ["test", "sample", "json"]
+  },
+  "config": {
+    "enabled": true,
+    "max_file_size": 1048576,
+    "supported_formats": [".txt", ".md", ".json", ".yaml", ".yml", ".csv", ".toml", ".pdf"]
+  }
+} 

test/unit/data/test_file_upload/sample.md

@@ -0,0 +1,20 @@
+# Sample Markdown File
+
+This is a sample markdown file for testing the file upload functionality.
+
+## Features
+
+- **Bold text** and *italic text*
+- Lists with bullets
+- Code blocks: `inline code`
+- Links: [Example](https://example.com)
+
+## Code Example
+
+```python
+def hello_world():
+    print("Hello, World!")
+    return "Success"
+```
+
+This file should be processed by the TXTFileUpload class since markdown is treated as text. 

test/unit/data/test_file_upload/sample.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+
+# Sample shell script for testing file upload functionality
+
+echo "Hello, World! This is a test script."
+
+# Function to demonstrate script functionality
+test_function() {
+    local message="$1"
+    echo "Testing: $message"
+    return 0
+}
+
+# Variables
+NAME="Test Script"
+VERSION="1.0.0"
+
+# Main execution
+echo "Running $NAME version $VERSION"
+
+# Test the function
+test_function "file upload functionality"
+
+# Conditional logic
+if [ -f "test_file.txt" ]; then
+    echo "Test file exists"
+else
+    echo "Test file does not exist"
+fi
+
+# Loop example
+for i in {1..3}; do
+    echo "Iteration $i"
+done
+
+echo "Script completed successfully!" 

test/unit/data/test_file_upload/sample.toml

@@ -0,0 +1,23 @@
+# Sample TOML configuration file
+name = "test_config"
+version = "1.0.0"
+description = "Sample TOML data for testing file upload functionality"
+
+[metadata]
+author = "Test User"
+created = "2024-01-01"
+tags = ["test", "sample", "toml"]
+
+[config]
+enabled = true
+max_file_size = 1048576
+supported_formats = [".txt", ".md", ".json", ".yaml", ".yml", ".csv", ".toml", ".pdf"]
+
+[features]
+text_processing = true
+file_upload = true
+chat_integration = true
+toml_support = true
+
+[nested.structure]
+with_deep_nesting = true 

test/unit/data/test_file_upload/sample.txt

@@ -0,0 +1,9 @@
+This is a sample text file for testing the file upload functionality.
+
+It contains multiple lines of text to test how the system handles:
+- Plain text content
+- Multiple lines
+- Special characters like: !@#$%^&*()
+- Numbers: 1234567890
+
+This file should be processed by the TXTFileUpload class. 

test/unit/data/test_file_upload/sample.yaml

@@ -0,0 +1,37 @@
+# Sample YAML configuration file
+name: test_config
+version: 1.0.0
+description: Sample YAML data for testing file upload functionality
+
+features:
+  - text_processing
+  - file_upload
+  - chat_integration
+  - yaml_support
+
+metadata:
+  author: Test User
+  created: 2024-01-01
+  tags:
+    - test
+    - sample
+    - yaml
+
+config:
+  enabled: true
+  max_file_size: 1048576
+  supported_formats:
+    - .txt
+    - .md
+    - .json
+    - .yaml
+    - .yml
+    - .csv
+    - .toml
+    - .pdf
+
+nested:
+  structure:
+    with:
+      deep:
+        nesting: true 

test/unit/test_config.py

@@ -153,7 +153,8 @@ def test_get_default_engine_with_toolboxenv(self):
     def test_get_default_engine_with_podman_available(self, platform, expected):
         with patch("ramalama.config.available", side_effect=lambda x: x == "podman"):
             with patch("sys.platform", platform):
-                assert get_default_engine() == expected
+                with patch("ramalama.config.apple_vm", return_value=False):
+                    assert get_default_engine() == expected
 
     def test_get_default_engine_with_podman_available_osx_apple_vm_has_podman(self):
         with patch("ramalama.config.available", side_effect=lambda x: x == "podman"):