Skip to content

Commit c4636da

Browse files
ruslanmvruslanmv
committed
First commit
1 parent 9623d69 commit c4636da

File tree

6 files changed

+52
-51
lines changed

6 files changed

+52
-51
lines changed

‎Makefile

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -98,6 +98,9 @@ docs: ## MkDocs live‑reload
9898
docs-build: ## Build static docs into site/
9999
mkdocs build
100100

101+
docs-deploy: ## Deploy documentation in production
102+
mkdocs gh-deploy --clean
103+
101104
# ----------------------------------------------------------------
102105
# Distribution
103106
# ----------------------------------------------------------------

‎README.md

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5,6 +5,7 @@
55
[![PyPI Version](https://img.shields.io/pypi/v/agent-generator.svg)](https://pypi.org/project/agent-generator/)
66
[![Python 3.9+](https://img.shields.io/pypi/pyversions/agent-generator.svg)](https://pypi.org/project/agent-generator/)
77
[![CI Status](https://github.com/ruslanmv/agent-generator/actions/workflows/ci.yml/badge.svg)](https://github.com/ruslanmv/agent-generator/actions/workflows/ci.yml)
8+
[![Documentation](https://img.shields.io/static/v1?label=docs&message=mkdocs&color=blue&logo=mkdocs)](https://ruslanmv.github.io/agent-generator/)
89
[![License Apache-2.0](https://img.shields.io/badge/license-Apache%202.0-blue)](https://github.com/ruslanmv/agent-generator/blob/master/LICENSE)
910

1011

@@ -191,6 +192,11 @@ make lint test
191192
mkdocs serve # live docs
192193
```
193194

195+
## Documentation
196+
197+
[https://ruslanmv.com/agent-generator/](https://ruslanmv.com/agent-generator/)
198+
199+
194200
---
195201

196202
## 🗺️ Roadmap

‎docs/architecture.md

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22

33
This document describes the overall architecture of **agent-generator**, with a focus on the generative pipeline that transforms plain‑English prompts into fully configured agent code or YAML for various frameworks.
44

5-
---
5+
66

77
## 1. High‑Level Overview
88

@@ -16,7 +16,7 @@ This document describes the overall architecture of **agent-generator**, with a
1616
8. **Framework Generator**: Another set of Jinja templates consume the `Workflow` and generate code or YAML in the target framework’s syntax, leveraging the LLM response only when needed (e.g., to fill details in templates).
1717
9. **Output**: The generated code/YAML is either printed with syntax highlighting or written to a file. Token‑ and cost‑estimates are optionally shown.
1818

19-
---
19+
2020

2121
## 2. Component Breakdown
2222

@@ -68,7 +68,7 @@ Supported frameworks:
6868
* **Syntax highlighting** via Rich when printing.
6969
* **MCP wrapper**: Optional FastAPI server scaffold if `--mcp` is passed.
7070

71-
---
71+
7272

7373
## 3. Generative Pipeline Detail
7474

@@ -95,7 +95,7 @@ flowchart TD
9595
* Decouples **what** you ask the LLM (open‑ended natural‑language) from **how** you produce deterministic code templates.
9696
* Future architectures can inject LLM responses deeper into code gen templates.
9797

98-
---
98+
9999

100100
## 4. Extensibility & Customization
101101

@@ -104,7 +104,7 @@ flowchart TD
104104
* **Custom Prompts**: Modify or add Jinja templates in `utils/prompts.py`.
105105
* **Parser tweaks**: Improve `utils/parser.py` to extract richer workflows (e.g. data schemas, error handling agents).
106106

107-
---
107+
108108

109109
## 5. File Locations
110110

‎docs/frameworks.md

Lines changed: 3 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -35,7 +35,7 @@ agent-generator [OPTIONS] "plain‑English requirement"
3535
| Cost estimate only | `agent-generator "Scrape website" -f react --dry-run --show-cost` |
3636
| Use **OpenAI** instead of WatsonX | `agent-generator "Write jokes" -f react -p openai --model gpt-4o` |
3737

38-
---
38+
3939

4040
## 2  Flask Web UI
4141

@@ -57,7 +57,6 @@ FLASK_APP=agent_generator.web FLASK_ENV=development flask run
5757

5858
![UI screenshot](images/ui-screenshot.png)
5959

60-
---
6160

6261
## 3  Docker usage
6362

@@ -74,7 +73,7 @@ You can also exec into the container to run the CLI:
7473
docker run --rm agent-generator agent-generator "Say hi" -f react --dry-run
7574
```
7675

77-
---
76+
7877

7978
## 4  Serving generated MCP skills
8079

@@ -87,7 +86,7 @@ python pipeline.py serve # exposes POST /invoke on port 8080
8786

8887
Upload the packaged script or its Docker image to your MCP Gateway and then **import** it as a custom skill in WatsonX Orchestrate.
8988

90-
---
89+
9190

9291
## 5  Troubleshooting
9392

‎docs/installation.md

Lines changed: 31 additions & 37 deletions
Original file line numberDiff line numberDiff line change
@@ -3,7 +3,6 @@
33

44
*agent-generator* supports **Python ≥ 3.9** and ships with IBM WatsonX as the default LLM provider. Optional extras unlock OpenAI support, the Flask Web UI, and developer tooling.
55

6-
-----
76

87
## Windows (via WSL)
98

@@ -12,32 +11,32 @@ If you’re on Windows, we recommend using WSL (Windows Subsystem for Linux) to
1211
**Enable WSL**
1312
Open PowerShell **as Administrator** and run:
1413

15-
```powershell
16-
wsl --install
17-
```
14+
```powershell
15+
wsl --install
16+
```
1817

1918
When it finishes, reboot your PC and launch your new Linux distro (e.g., Ubuntu).
2019

2120
**Create & activate a virtual environment**
2221
In your WSL terminal:
2322

24-
```bash
25-
python3 -m venv venv
26-
source venv/bin/activate
27-
```
23+
```bash
24+
python3 -m venv venv
25+
source venv/bin/activate
26+
```
2827

2928
**Upgrade pip & install agent-generator**
3029

31-
```bash
32-
pip install --upgrade pip --break-system-packages
33-
pip install "agent-generator[dev,web,openai]" --break-system-packages
34-
```
30+
```bash
31+
pip install --upgrade pip --break-system-packages
32+
pip install "agent-generator[dev,web,openai]" --break-system-packages
33+
```
3534

36-
If you encounter PEP 668 “externally managed environment” errors inside the `venv`, continue using the `--break-system-packages` flag.
35+
If you encounter PEP 668 “externally managed environment” errors inside the `venv`, continue using the `--break-system-packages` flag.
3736

3837
**Prepare the `.env` file**
3938

40-
At the project root (next to `Makefile`, `pyproject.toml`, etc.), create a file named `.env`:
39+
At the project root (next to `Makefile`, `pyproject.toml`, etc.), create a file named `.env`:
4140

4241
```bash
4342
WATSONX_API_KEY=your_watsonx_key
@@ -57,7 +56,7 @@ set -a && source .env && set +a
5756
# Method B: Strip Windows carriage returns and export
5857
export $(cat .env | tr -d '\r' | xargs)
5958
```
60-
Verify that the variables were loaded correctly:
59+
Verify that the variables were loaded correctly:
6160

6261
```bash
6362
echo "$WATSONX_API_KEY"
@@ -67,23 +66,23 @@ echo "$WATSONX_PROJECT_ID"
6766
**Run the generator**
6867
To generate an agent using the default WatsonX provider:
6968

70-
```bash
71-
agent-generator \
72-
"I need a research assistant that summarises papers" \
73-
--framework watsonx_orchestrate \
74-
--output research_assistant.yaml
75-
```
69+
```bash
70+
agent-generator \
71+
"I need a research assistant that summarises papers" \
72+
--framework watsonx_orchestrate \
73+
--output research_assistant.yaml
74+
```
7675

7776
Or, to use the OpenAI provider:
7877

79-
```bash
80-
export OPENAI_API_KEY=sk-...
81-
agent-generator \
82-
"I need a research assistant that summarises papers" \
83-
--framework crewai \
84-
--provider openai \
85-
--output research_assistant.py
86-
```
78+
```bash
79+
export OPENAI_API_KEY=sk-...
80+
agent-generator \
81+
"I need a research assistant that summarises papers" \
82+
--framework crewai \
83+
--provider openai \
84+
--output research_assistant.py
85+
```
8786

8887
## Basic Installation
8988

@@ -95,9 +94,9 @@ pip install agent-generator
9594

9695
This gives you:
9796

98-
* The CLI (`agent-generator …`)
99-
* Core runtime dependencies
100-
* The WatsonX provider (default model: `meta-llama/llama-3-3-70b-instruct`)
97+
* The CLI (`agent-generator …`)
98+
* Core runtime dependencies
99+
* The WatsonX provider (default model: `meta-llama/llama-3-3-70b-instruct`)
101100

102101
-----
103102

@@ -170,8 +169,6 @@ import json
170169
...
171170
```
172171

173-
-----
174-
175172
## Running the Web UI
176173

177174
You can run the web interface using either the Flask development server or Docker.
@@ -193,7 +190,6 @@ docker run -e WATSONX_API_KEY=... -p 8000:8000 agentgen
193190

194191
Access the UI at `http://localhost:8000`.
195192

196-
-----
197193

198194
## Upgrading
199195

@@ -209,7 +205,6 @@ pip install --upgrade agent-generator
209205
> pip install --upgrade "agent-generator[web,openai]"
210206
> ```
211207
212-
-----
213208
214209
## Troubleshooting
215210
@@ -221,5 +216,4 @@ pip install --upgrade agent-generator
221216
| Mermaid diagram not rendering (UI) | Ensure your browser has internet access to the CDN at `unpkg.com`. |
222217
223218
224-
225219
Jump in: **[Installation ➜](installation.md)** · **[Usage ➜](usage.md)** · **[Frameworks ➜](frameworks.md)**

‎docs/usage.md

Lines changed: 4 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -35,7 +35,7 @@ agent-generator [OPTIONS] "plain‑English requirement"
3535
| Cost estimate only | `agent-generator "Scrape website" -f react --dry-run --show-cost` |
3636
| Use **OpenAI** instead of WatsonX | `agent-generator "Write jokes" -f react -p openai --model gpt-4o` |
3737

38-
---
38+
3939

4040
## 2  Flask Web UI
4141

@@ -57,7 +57,7 @@ FLASK_APP=agent_generator.web FLASK_ENV=development flask run
5757

5858
![UI screenshot](images/ui-screenshot.png)
5959

60-
---
60+
6161

6262
## 3  Docker usage
6363

@@ -74,7 +74,7 @@ You can also exec into the container to run the CLI:
7474
docker run --rm agent-generator agent-generator "Say hi" -f react --dry-run
7575
```
7676

77-
---
77+
7878

7979
## 4  Serving generated MCP skills
8080

@@ -87,7 +87,7 @@ python pipeline.py serve # exposes POST /invoke on port 8080
8787

8888
Upload the packaged script or its Docker image to your MCP Gateway and then **import** it as a custom skill in WatsonX Orchestrate.
8989

90-
---
90+
9191

9292
## 5  Troubleshooting
9393

@@ -101,7 +101,6 @@ Upload the packaged script or its Docker image to your MCP Gateway and then **i
101101

102102
Still stuck? Open an issue on the [GitHub tracker](https://github.com/ruslanmv/agent-generator/issues).
103103

104-
---
105104

106105
Jump in: **[Installation ➜](installation.md)** · **[Usage ➜](usage.md)** · **[Frameworks ➜](frameworks.md)**
107106

0 commit comments

Comments
 (0)