https://github.com/shibing624/codeassist
CodeAssist is an advanced code completion tool that provides high-quality code completions for Python, Java, C++ and so on. CodeAssist 是一个高级代码补全工具，高质量为 Python、Java 和 C++ 补全代码。
https://github.com/shibing624/codeassist
auto-completion code-autocomplete code-generation gpt-4 gpt2 starcoder wizardcoder
Last synced: 5 months ago
JSON representation
CodeAssist is an advanced code completion tool that provides high-quality code completions for Python, Java, C++ and so on. CodeAssist 是一个高级代码补全工具，高质量为 Python、Java 和 C++ 补全代码。
- Host: GitHub
- URL: https://github.com/shibing624/codeassist
- Owner: shibing624
- License: apache-2.0
- Created: 2022-02-09T08:20:08.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2024-02-19T07:41:35.000Z (over 1 year ago)
- Last Synced: 2025-04-21T06:57:57.565Z (6 months ago)
- Topics: auto-completion, code-autocomplete, code-generation, gpt-4, gpt2, starcoder, wizardcoder
- Language: Python
- Homepage:
- Size: 1.06 MB
- Stars: 58
- Watchers: 3
- Forks: 8
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Citation: CITATION.cff
Awesome Lists containing this project
README
[**🇨🇳中文**](https://github.com/shibing624/codeassist/blob/main/README.md) | [**🌐English**](https://github.com/shibing624/codeassist/blob/main/README_EN.md) | [**📖文档/Docs**](https://github.com/shibing624/codeassist/wiki) | [**🤖模型/Models**](https://huggingface.co/shibing624)
-----------------
# CodeAssist: Advanced Code Completion Tool
[](https://badge.fury.io/py/CodeAssist)
[](CONTRIBUTING.md)
[](https://github.com/shibing624/CodeAssist/graphs/contributors)
[](LICENSE)
[](requirements.txt)
[](https://github.com/shibing624/CodeAssist/issues)
[](#Contact)
## Introduction
**CodeAssist** is an advanced code completion tool that intelligently provides high-quality code completions for Python, Java, and C++ and so on.
CodeAssist 是一个高级代码补全工具，高质量为 Python、Java 和 C++ 等编程语言补全代码
## Features
- GPT based code completion
- Code completion for `Python`, `Java`, `C++`, `javascript` and so on
- Line and block code completion
- Train(Fine-tuning) and predict model with your own data
### Release Models
| Arch | BaseModel | Model | Model Size |
|:-------|:------------------|:------------------------------------------------------------------------------------------------------------------------|:----------:|
| GPT | gpt2 | [shibing624/code-autocomplete-gpt2-base](https://huggingface.co/shibing624/code-autocomplete-gpt2-base) | 487MB |
| GPT | distilgpt2 | [shibing624/code-autocomplete-distilgpt2-python](https://huggingface.co/shibing624/code-autocomplete-distilgpt2-python) | 319MB |
| GPT | bigcode/starcoder | [WizardLM/WizardCoder-15B-V1.0](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0) | 29GB |
### Demo
HuggingFace Demo: https://huggingface.co/spaces/shibing624/code-autocomplete
backend model: `shibing624/code-autocomplete-gpt2-base`
## Install
```shell
pip install torch # conda install pytorch
pip install -U codeassist
```
or
```shell
git clone https://github.com/shibing624/codeassist.git
cd CodeAssist
python setup.py install
```
## Usage
### WizardCoder model
WizardCoder-15b is fine-tuned `bigcode/starcoder` with alpaca code data, you can use the following code to generate code:
example: [examples/wizardcoder_demo.py](https://github.com/shibing624/CodeAssist/blob/main/examples/wizardcoder_demo.py)
```python
import sys
sys.path.append('..')
from codeassist import WizardCoder
m = WizardCoder("WizardLM/WizardCoder-15B-V1.0")
print(m.generate('def load_csv_file(file_path):')[0])
```
output:
```python
import csv
def load_csv_file(file_path):
"""
Load data from a CSV file and return a list of dictionaries.
"""
# Open the file in read mode
with open(file_path, 'r') as file:
# Create a CSV reader object
csv_reader = csv.DictReader(file)
# Initialize an empty list to store the data
data = []
# Iterate over each row of data
for row in csv_reader:
# Append the row of data to the list
data.append(row)
# Return the list of data
return data
```
model output is impressively effective, it currently supports English and Chinese input, you can enter instructions or code prefixes as required.
### distilgpt2 model
distilgpt2 fine-tuned code autocomplete model, you can use the following code:
example: [examples/distilgpt2_demo.py](https://github.com/shibing624/CodeAssist/blob/main/examples/distilgpt2_demo.py)
```python
import sys
sys.path.append('..')
from codeassist import GPT2Coder
m = GPT2Coder("shibing624/code-autocomplete-distilgpt2-python")
print(m.generate('import torch.nn as')[0])
```
output:
```shell
import torch.nn as nn
import torch.nn.functional as F
```
### Use with huggingface/transformers：
example: [examples/use_transformers_gpt2.py](https://github.com/shibing624/CodeAssist/blob/main/examples/use_transformers_gpt2.py)
### Train Model
#### Train WizardCoder model
example: [examples/training_wizardcoder_mydata.py](https://github.com/shibing624/CodeAssist/blob/main/examples/training_wizardcoder_mydata.py)
```shell
cd examples
CUDA_VISIBLE_DEVICES=0,1 python training_wizardcoder_mydata.py --do_train --do_predict --num_epochs 1 --output_dir outputs-wizard --model_name WizardLM/WizardCoder-15B-V1.0
```
- GPU memory: 31GB
- finetune need 2*V100(32GB)
- inference need 1*V100(32GB)
#### Train distilgpt2 model
example: [examples/training_gpt2_mydata.py](https://github.com/shibing624/CodeAssist/blob/main/examples/training_gpt2_mydata.py)
```shell
cd examples
python training_gpt2_mydata.py --do_train --do_predict --num_epochs 15 --output_dir outputs-gpt2 --model_name gpt2
```
PS: fine-tuned result model is GPT2-python: [shibing624/code-autocomplete-gpt2-base](https://huggingface.co/shibing624/code-autocomplete-gpt2-base),
I spent about 24 hours with V100 to fine-tune it.
### Server
start FastAPI server:
example: [examples/server.py](https://github.com/shibing624/CodeAssist/blob/main/examples/server.py)
```shell
cd examples
python server.py
```
open url: http://0.0.0.0:8001/docs
## Dataset
This allows to customize dataset building. Below is an example of the building process.
Let's use Python codes from [Awesome-pytorch-list](https://github.com/bharathgs/Awesome-pytorch-list)
1. We want the model to help auto-complete codes at a general level. The codes of The Algorithms suits the need.
2. This code from this project is well written (high-quality codes).
dataset tree:
```shell
examples/download/python
├── train.txt
└── valid.txt
└── test.txt
```
There are three ways to build dataset:
1. Use the huggingface/datasets library load the dataset
huggingface datasets [https://huggingface.co/datasets/shibing624/source_code](https://huggingface.co/datasets/shibing624/source_code)
```python
from datasets import load_dataset
dataset = load_dataset("shibing624/source_code", "python") # python or java or cpp
print(dataset)
print(dataset['test'][0:10])
```
output:
```shell
DatasetDict({
train: Dataset({
features: ['text'],
num_rows: 5215412
})
validation: Dataset({
features: ['text'],
num_rows: 10000
})
test: Dataset({
features: ['text'],
num_rows: 10000
})
})
{'text': [
" {'max_epochs': [1, 2]},\n",
' refit=False,\n', ' cv=3,\n',
" scoring='roc_auc',\n", ' )\n',
' search.fit(*data)\n',
'',
' def test_module_output_not_1d(self, net_cls, data):\n',
' from skorch.toy import make_classifier\n',
' module = make_classifier(\n'
]}
```
2. Download dataset from Cloud
| Name | Source | Download | Size |
| :------- | :--------- | :---------: | :---------: |
| Python+Java+CPP source code | Awesome-pytorch-list(5.22 Million lines) | [github_source_code.zip](https://github.com/shibing624/codeassist/releases/download/0.0.4/source_code.zip) | 105M |
download dataset and unzip it, put to `examples/`.
3. Get source code from scratch and build dataset
[prepare_code_data.py](https://github.com/shibing624/CodeAssist/blob/main/examples/prepare_code_data.py)
```shell
cd examples
python prepare_code_data.py --num_repos 260
```
## Contact
- Issue(建议)
：[](https://github.com/shibing624/CodeAssist/issues)
- 邮件我：xuming: xuming624@qq.com
- 微信我： 加我*微信号：xuming624, 备注：个人名称-公司-NLP* 进NLP交流群。
## Citation
如果你在研究中使用了codeassist，请按如下格式引用：
APA:
```latex
Xu, M. codeassist: Code AutoComplete with GPT model (Version 1.0.0) [Computer software]. https://github.com/shibing624/codeassist
```
BibTeX:
```latex
@software{Xu_codeassist,
author = {Ming Xu},
title = {CodeAssist: Code AutoComplete with Generation model},
url = {https://github.com/shibing624/codeassist},
version = {1.0.0}
}
```
## License
This repository is licensed under the [The Apache License 2.0](LICENSE).
Please follow the [Attribution-NonCommercial 4.0 International](https://github.com/nlpxucan/WizardLM/blob/main/WizardCoder/MODEL_WEIGHTS_LICENSE) to use the WizardCoder model.
## Contribute
项目代码还很粗糙，如果大家对代码有所改进，欢迎提交回本项目，在提交之前，注意以下两点：
- 在`tests`添加相应的单元测试
- 使用`python setup.py test`来运行所有单元测试，确保所有单测都是通过的
之后即可提交PR。
## Reference
- [gpt-2-simple](https://github.com/minimaxir/gpt-2-simple)
- [galois-autocompleter](https://github.com/galois-autocompleter/galois-autocompleter)
- [WizardLM/WizardCoder-15B-V1.0](https://huggingface.co/WizardLM/WizardCoder-15B-V1.0)