| Name | prompt-gear JSON |
| Version |
0.1.0b2
JSON |
| download |
| home_page | None |
| Summary | YAML-powered prompt manager with multi-backend support |
| upload_time | 2025-07-23 01:08:28 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.9 |
| license | MIT |
| keywords |
ai
llm
management
prompt
template
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# Prompt Gear
🧠 **YAML-powered prompt manager with multi-backend support**
Prompt Gear is a flexible prompt management tool that can be embedded into any Python application (LangGraph, LangChain, FastAPI, etc.). It supports multiple storage backends and provides both CLI and Python SDK interfaces.
## ✨ Features
- ✅ **Multi-backend storage**: Filesystem (YAML), SQLite, PostgreSQL
- ✅ **YAML-based prompt format**: Human-readable and version-controllable
- ✅ **File input support**: Create prompts from files for better organization
- ✅ **Advanced version management**: Sequence-based versioning with explicit latest tracking
- ✅ **Dual interface**: CLI tool + Python SDK
- ✅ **Flexible configuration**: System prompts, user prompts, and custom config
- ✅ **Environment-based config**: Use `.env` files for backend configuration
- ✅ **Robust data integrity**: Transaction-based operations with automatic rollback
- ✅ **uv + pyproject.toml**: Modern Python packaging ready for PyPI
## 🚀 Quick Start
### Installation
```bash
pip install prompt-gear
```
### Initialize in your project
```bash
# Initialize with filesystem backend (default)
promptgear init --backend filesystem
# Initialize with SQLite backend
promptgear init --backend sqlite
# Initialize with PostgreSQL backend
promptgear init --backend postgres
```
### Create a prompt
```bash
# Direct input
promptgear create chatbot_greeting --version v1 \
--system "You are a helpful assistant that speaks politely." \
--user "Hello! How can I help you? {{user_input}}" \
--config '{"temperature": 0.7, "max_tokens": 512}'
# File input for complex prompts
promptgear create complex_chatbot \
--system-file system_prompt.txt \
--user-file user_template.txt \
--config-file config.yaml
# Mix file and direct input
promptgear create hybrid_prompt \
--system-file detailed_system.txt \
--user "Simple user message: {{input}}" \
--config '{"temperature": 0.5}'
```
### Use in Python
```python
from promptgear import PromptManager
pm = PromptManager()
prompt = pm.get_prompt("chatbot_greeting", "v1")
print(prompt.system_prompt)
print(prompt.user_prompt)
print(prompt.config["temperature"])
```
## 📁 YAML Format
Prompts are stored in YAML format:
```yaml
name: chatbot_greeting
version: v1
system_prompt: >
You are a helpful assistant that speaks politely.
user_prompt: >
Hello! How can I help you? {{user_input}}
config:
temperature: 0.7
max_tokens: 512
```
## 🔧 Configuration
Configure via `.env` file:
```env
# Filesystem backend
PROMPT_GEAR_BACKEND=filesystem
PROMPT_GEAR_PROMPT_DIR=./prompts
# SQLite backend
PROMPT_GEAR_BACKEND=sqlite
PROMPT_GEAR_DB_URL=sqlite:///prompts.db
# PostgreSQL backend
PROMPT_GEAR_BACKEND=postgres
PROMPT_GEAR_DB_URL=postgresql://user:pass@localhost/prompts
```
## � Version Management
Prompt Gear uses an advanced sequence-based versioning system:
### Version Sequence Numbers
- Each prompt version gets an auto-incrementing sequence number
- Sequence numbers are maintained independently per prompt name
- Deleted versions don't affect sequence numbering (e.g., v1→v2→v4 after deleting v3)
### Latest Version Tracking
- Each prompt has exactly one version marked as "latest"
- Creating a new version automatically updates the latest flag
- Deleting the latest version automatically promotes the highest sequence number version
### Examples
```python
# Create versions - sequence numbers auto-assigned
v1 = pm.create_prompt("example", "v1", "System v1", "User v1") # seq=1, latest=True
v2 = pm.create_prompt("example", "v2", "System v2", "User v2") # seq=2, latest=True
# v1 is automatically marked as not latest
v1_updated = pm.get_prompt("example", "v1") # seq=1, latest=False
# Get latest version without specifying version
latest = pm.get_prompt("example") # Returns v2
# Delete latest version
pm.delete_prompt("example", "v2")
new_latest = pm.get_prompt("example") # Returns v1 (automatically promoted)
```
## �📚 CLI Commands
- `promptgear init` - Initialize Prompt Gear with backend selection
- `promptgear create` - Create a new prompt (automatically becomes latest)
- `promptgear get` - Get a prompt (latest version if version not specified)
- `promptgear list` - List all prompts with their latest versions
- `promptgear delete` - Delete a prompt version (auto-promotes new latest)
- `promptgear versions` - List all versions of a prompt
- `promptgear status` - Show backend status and statistics
## 🧩 Python SDK
```python
from promptgear import PromptManager
# Initialize
pm = PromptManager()
# Create prompt (automatically gets sequence number and latest flag)
prompt = pm.create_prompt(
name="my_prompt",
version="v1",
system_prompt="You are helpful.",
user_prompt="{{user_input}}",
config={"temperature": 0.8}
)
# Get latest version (no version specified)
latest = pm.get_prompt("my_prompt")
# Get specific version
specific = pm.get_prompt("my_prompt", "v1")
# List prompts (returns latest versions)
prompts = pm.list_prompts()
# List all versions of a prompt
versions = pm.list_versions("my_prompt")
# Update prompt
pm.update_prompt("my_prompt", "v1", system_prompt="Updated system prompt")
# Delete prompt version
pm.delete_prompt("my_prompt", "v1") # Latest flag auto-managed
```
## 🏗️ Development Status
Currently in **Phase 3** development:
- ✅ Basic schema and filesystem backend
- ✅ Core PromptManager functionality
- ✅ CLI interface with backend selection
- ✅ SQLite backend with structured storage
- ✅ PostgreSQL backend with connection pooling
- ✅ Advanced version management with sequence numbers
- ✅ Robust data integrity and transaction support
- ✅ Comprehensive test coverage across all backends
- ⏳ Advanced features and optimizations (Phase 4)
- ⏳ PyPI packaging (Phase 4)
## 📄 License
MIT License
Raw data
{
"_id": null,
"home_page": null,
"name": "prompt-gear",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "AI, LLM, management, prompt, template",
"author": null,
"author_email": "miniGears <wenyi@wygears.com>",
"download_url": "https://files.pythonhosted.org/packages/26/36/819610cba5e113a23f50bce7cf02a529ed13fc355557471733c835198c2b/prompt_gear-0.1.0b2.tar.gz",
"platform": null,
"description": "# Prompt Gear\n\n\ud83e\udde0 **YAML-powered prompt manager with multi-backend support**\n\nPrompt Gear is a flexible prompt management tool that can be embedded into any Python application (LangGraph, LangChain, FastAPI, etc.). It supports multiple storage backends and provides both CLI and Python SDK interfaces.\n\n## \u2728 Features\n\n- \u2705 **Multi-backend storage**: Filesystem (YAML), SQLite, PostgreSQL\n- \u2705 **YAML-based prompt format**: Human-readable and version-controllable\n- \u2705 **File input support**: Create prompts from files for better organization\n- \u2705 **Advanced version management**: Sequence-based versioning with explicit latest tracking\n- \u2705 **Dual interface**: CLI tool + Python SDK\n- \u2705 **Flexible configuration**: System prompts, user prompts, and custom config\n- \u2705 **Environment-based config**: Use `.env` files for backend configuration\n- \u2705 **Robust data integrity**: Transaction-based operations with automatic rollback\n- \u2705 **uv + pyproject.toml**: Modern Python packaging ready for PyPI\n\n## \ud83d\ude80 Quick Start\n\n### Installation\n\n```bash\npip install prompt-gear\n```\n\n### Initialize in your project\n\n```bash\n# Initialize with filesystem backend (default)\npromptgear init --backend filesystem\n\n# Initialize with SQLite backend\npromptgear init --backend sqlite\n\n# Initialize with PostgreSQL backend\npromptgear init --backend postgres\n```\n\n### Create a prompt\n\n```bash\n# Direct input\npromptgear create chatbot_greeting --version v1 \\\n --system \"You are a helpful assistant that speaks politely.\" \\\n --user \"Hello! How can I help you? {{user_input}}\" \\\n --config '{\"temperature\": 0.7, \"max_tokens\": 512}'\n\n# File input for complex prompts\npromptgear create complex_chatbot \\\n --system-file system_prompt.txt \\\n --user-file user_template.txt \\\n --config-file config.yaml\n\n# Mix file and direct input\npromptgear create hybrid_prompt \\\n --system-file detailed_system.txt \\\n --user \"Simple user message: {{input}}\" \\\n --config '{\"temperature\": 0.5}'\n```\n\n### Use in Python\n\n```python\nfrom promptgear import PromptManager\n\npm = PromptManager()\nprompt = pm.get_prompt(\"chatbot_greeting\", \"v1\")\n\nprint(prompt.system_prompt)\nprint(prompt.user_prompt)\nprint(prompt.config[\"temperature\"])\n```\n\n## \ud83d\udcc1 YAML Format\n\nPrompts are stored in YAML format:\n\n```yaml\nname: chatbot_greeting\nversion: v1\nsystem_prompt: >\n You are a helpful assistant that speaks politely.\nuser_prompt: >\n Hello! How can I help you? {{user_input}}\nconfig:\n temperature: 0.7\n max_tokens: 512\n```\n\n## \ud83d\udd27 Configuration\n\nConfigure via `.env` file:\n\n```env\n# Filesystem backend\nPROMPT_GEAR_BACKEND=filesystem\nPROMPT_GEAR_PROMPT_DIR=./prompts\n\n# SQLite backend\nPROMPT_GEAR_BACKEND=sqlite\nPROMPT_GEAR_DB_URL=sqlite:///prompts.db\n\n# PostgreSQL backend\nPROMPT_GEAR_BACKEND=postgres\nPROMPT_GEAR_DB_URL=postgresql://user:pass@localhost/prompts\n```\n\n## \ufffd Version Management\n\nPrompt Gear uses an advanced sequence-based versioning system:\n\n### Version Sequence Numbers\n- Each prompt version gets an auto-incrementing sequence number\n- Sequence numbers are maintained independently per prompt name\n- Deleted versions don't affect sequence numbering (e.g., v1\u2192v2\u2192v4 after deleting v3)\n\n### Latest Version Tracking\n- Each prompt has exactly one version marked as \"latest\"\n- Creating a new version automatically updates the latest flag\n- Deleting the latest version automatically promotes the highest sequence number version\n\n### Examples\n\n```python\n# Create versions - sequence numbers auto-assigned\nv1 = pm.create_prompt(\"example\", \"v1\", \"System v1\", \"User v1\") # seq=1, latest=True\nv2 = pm.create_prompt(\"example\", \"v2\", \"System v2\", \"User v2\") # seq=2, latest=True\n\n# v1 is automatically marked as not latest\nv1_updated = pm.get_prompt(\"example\", \"v1\") # seq=1, latest=False\n\n# Get latest version without specifying version\nlatest = pm.get_prompt(\"example\") # Returns v2\n\n# Delete latest version\npm.delete_prompt(\"example\", \"v2\")\nnew_latest = pm.get_prompt(\"example\") # Returns v1 (automatically promoted)\n```\n\n## \ufffd\ud83d\udcda CLI Commands\n\n- `promptgear init` - Initialize Prompt Gear with backend selection\n- `promptgear create` - Create a new prompt (automatically becomes latest)\n- `promptgear get` - Get a prompt (latest version if version not specified)\n- `promptgear list` - List all prompts with their latest versions\n- `promptgear delete` - Delete a prompt version (auto-promotes new latest)\n- `promptgear versions` - List all versions of a prompt\n- `promptgear status` - Show backend status and statistics\n\n## \ud83e\udde9 Python SDK\n\n```python\nfrom promptgear import PromptManager\n\n# Initialize\npm = PromptManager()\n\n# Create prompt (automatically gets sequence number and latest flag)\nprompt = pm.create_prompt(\n name=\"my_prompt\",\n version=\"v1\",\n system_prompt=\"You are helpful.\",\n user_prompt=\"{{user_input}}\",\n config={\"temperature\": 0.8}\n)\n\n# Get latest version (no version specified)\nlatest = pm.get_prompt(\"my_prompt\")\n\n# Get specific version\nspecific = pm.get_prompt(\"my_prompt\", \"v1\")\n\n# List prompts (returns latest versions)\nprompts = pm.list_prompts()\n\n# List all versions of a prompt\nversions = pm.list_versions(\"my_prompt\")\n\n# Update prompt\npm.update_prompt(\"my_prompt\", \"v1\", system_prompt=\"Updated system prompt\")\n\n# Delete prompt version\npm.delete_prompt(\"my_prompt\", \"v1\") # Latest flag auto-managed\n```\n\n## \ud83c\udfd7\ufe0f Development Status\n\nCurrently in **Phase 3** development:\n- \u2705 Basic schema and filesystem backend\n- \u2705 Core PromptManager functionality\n- \u2705 CLI interface with backend selection\n- \u2705 SQLite backend with structured storage\n- \u2705 PostgreSQL backend with connection pooling\n- \u2705 Advanced version management with sequence numbers\n- \u2705 Robust data integrity and transaction support\n- \u2705 Comprehensive test coverage across all backends\n- \u23f3 Advanced features and optimizations (Phase 4)\n- \u23f3 PyPI packaging (Phase 4)\n\n## \ud83d\udcc4 License\n\nMIT License\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "YAML-powered prompt manager with multi-backend support",
"version": "0.1.0b2",
"project_urls": {
"Documentation": "https://github.com/miniGears/prompt-gear#readme",
"Homepage": "https://github.com/miniGears/prompt-gear",
"Issues": "https://github.com/miniGears/prompt-gear/issues",
"Repository": "https://github.com/miniGears/prompt-gear"
},
"split_keywords": [
"ai",
" llm",
" management",
" prompt",
" template"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "0eb861da4e892683f752577488851c956b20ccb0d2c9a3b5d127754b59f5ef10",
"md5": "ea1bc8ab0cfbcf60c6809614c85184d8",
"sha256": "d918ce53a7855c8412146c7100058045e4cdf30307a9e2a88c7adf1d083fb7cb"
},
"downloads": -1,
"filename": "prompt_gear-0.1.0b2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "ea1bc8ab0cfbcf60c6809614c85184d8",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 26698,
"upload_time": "2025-07-23T01:08:26",
"upload_time_iso_8601": "2025-07-23T01:08:26.713006Z",
"url": "https://files.pythonhosted.org/packages/0e/b8/61da4e892683f752577488851c956b20ccb0d2c9a3b5d127754b59f5ef10/prompt_gear-0.1.0b2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "2636819610cba5e113a23f50bce7cf02a529ed13fc355557471733c835198c2b",
"md5": "089de1dc46bd8999a112b11b4bc87e09",
"sha256": "7051f51882dc3e2c4c1f8ac85e214aa0cf4c896cb8a21054f970f4bfd9bf3f02"
},
"downloads": -1,
"filename": "prompt_gear-0.1.0b2.tar.gz",
"has_sig": false,
"md5_digest": "089de1dc46bd8999a112b11b4bc87e09",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 69099,
"upload_time": "2025-07-23T01:08:28",
"upload_time_iso_8601": "2025-07-23T01:08:28.123964Z",
"url": "https://files.pythonhosted.org/packages/26/36/819610cba5e113a23f50bce7cf02a529ed13fc355557471733c835198c2b/prompt_gear-0.1.0b2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-07-23 01:08:28",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "miniGears",
"github_project": "prompt-gear#readme",
"github_not_found": true,
"lcname": "prompt-gear"
}
