# MatchBox
Industrial recommender systems generally have two main stages: matching and ranking. In the first stage, candidate item matching (also known as candidate retrieval) aims for efficient and high-recall retrieval from a large item corpus. MatchBox provides an open source library for candidate item matching, with stunning features in configurability, tunability, and reproducibility.
## Model Zoo
| Publication | Model | Paper | Benchmark |
|:-----------:|:--------------:|:----------------------------------------------------------------- |:-------------:|
| UAI'09 | [MF-BPR](./model_zoo/MF) | [BPR: Bayesian Personalized Ranking from Implicit Feedback](https://arxiv.org/ftp/arxiv/papers/1205/1205.2618.pdf) | [:arrow_upper_right:](./model_zoo/MF/config) |
| RecSys'16 | [YoutubeNet](./model_zoo/YoutubeNet) | [Deep Neural Networks for YouTube Recommendations](https://dl.acm.org/doi/10.1145/2959100.2959190) | [:arrow_upper_right:](./model_zoo/YouTubeNet/config) |
| CIKM'21 | [MF-CCL](./model_zoo/MF)/ [SimpleX](./model_zoo/SimpleX) | [SimpleX: A Simple and Strong Baseline for Collaborative Filtering](https://arxiv.org/abs/2109.12613) | [:arrow_upper_right:](./model_zoo/SimpleX/config) |
## Dependency
We suggest to use the following environment where we test MatchBox only.
+ CUDA 10.0
+ python 3.6
+ pytorch 1.0
+ PyYAML
+ pandas
+ scikit-learn
+ numpy
+ h5py
+ tqdm
## Get Started
The code workflow is structured as follows:
```python
# Set the dataset config and model config
feature_cols = [{...}] # define feature columns
label_col = {...} # define label column
params = {...} # set data params and model params
# Set the feature encoding specs
feature_encoder = FeatureEncoder(feature_cols, label_col, ...) # define the feature encoder
datasets.build_dataset(feature_encoder, ...) # fit feature_encoder and build dataset
# Load data generators
train_gen, valid_gen, test_gen = datasets.h5_generator(feature_encoder, ...)
# Define a model
model = SimpleX(...)
# Train the model
model.fit(train_gen, valid_gen, ...)
# Evaluation
model.evaluate(test_gen)
```
#### Run the code
For reproducing the experiment results, you can run the benchmarking scripts with the corresponding configs as follows.
+ --config: The config directory where dataset config and model config are located.
+ --expid: The experiment id defined in a model config file to set a group of hyper-parameters.
+ --gpu: The gpu index used for experiment, and -1 for CPU.
```bash
cd data/Yelp18/yelp18_m1
python matchbox_convert_data.py
cd model_zoo/SimpleX
python run_expid.py --config ./config/SimpleX_yelp18_m1 --expid SimpleX_yelp18_m1 --gpu 0
...
python run_expid.py --config ./config/SimpleX_amazonbooks_m1 --expid SimpleX_amazonbooks_m1 --gpu 0
python run_expid.py --config ./config/SimpleX_gowalla_m1 --expid SimpleX_gowalla_m1 --gpu 0
```
The running logs are also available in each config directory.
用于候选项目匹配的可配置、可调和可重现的库_Python_下载.zip （87个子文件） 
MatchBox-main model_zoo 
__init__.py 104B YouTubeNet src __init__.py 34B YouTubeNet.py 4KB run_expid.py 3KB config YouTubeNet_yelp18_m1 dataset_config.yaml 1020B model_config.yaml 950B SimpleX src __init__.py 28B SimpleX.py 8KB run_expid.py 3KB config SimpleX_gowalla_m1 dataset_config.yaml 954B SimpleX_gowalla_m1_013_4ecb0cbe.log 35KB model_config.yaml 1008B SimpleX_amazonbooks_m1 SimpleX_amazonbooks_m1_003_a30a8992.log 33KB dataset_config.yaml 1KB model_config.yaml 1KB SimpleX_yelp18_m1 dataset_config.yaml 997B model_config.yaml 995B SimpleX_yelp18_m1_034_297a4b82.log 38KB MF src __init__.py 46B MF.py 2KB BiasMF.py 4KB run_expid.py 3KB config MF_CCL_yelp18_m1 dataset_config.yaml 793B model_config.yaml 778B MF_BPR_yelp18_m1 dataset_config.yaml 793B model_config.yaml 772B benchmark run_expid_list.py 771B enumerate_expid_list.py 985B run_expid.py 3KB run_param_tuner.py 793B data Yelp18 yelp18_m1 matchbox_convert_data.py 2KB
test.txt 1.89MB README.md 298B train.txt 6.59MB README.md 136B LICENSE 11KB matchbox utils.py 4KB __init__.py 32B ann __init__.py 43B faiss.py 614B preprocess.py 5KB version.py 23B features.py 17KB metrics.py 6KB autotuner.py 8KB datasets __init__.py 25B data_utils.py 7KB pytorch __init__.py 0B layers __init__.py 67B embedding.py 7KB sequence.py 660B mlp.py 2KB __pycache__ embedding.cpython-36.pyc 4KB sequence.cpython-36.pyc 1KB mlp.cpython-36.pyc 1KB __init__.cpython-36.pyc 178B data_generator.py 10KB losses __init__.py 324B pairwise_logistic_loss.py 548B cosine_contrastive_loss.py 1KB sigmoid_crossentropy_loss.py 605B softmax_crossentropy_loss.py 572B pairwise_margin_loss.py 611B mse_loss.py 580B __pycache__ pairwise_logistic_loss.cpython-36.pyc 1000B pairwise_margin_loss.cpython-36.pyc 1KB sigmoid_crossentropy_loss.cpython-36.pyc 1KB softmax_crossentropy_loss.cpython-36.pyc 1KB mse_loss.cpython-36.pyc 978B __init__.cpython-36.pyc 502B cosine_contrastive_loss.cpython-36.pyc 1KB models __init__.py 34B base_model.py 11KB __pycache__ MF.cpython-36.pyc 2KB YoutubeDNN.cpython-36.pyc 3KB __init__.cpython-36.pyc 205B SimpleX.cpython-36.pyc 5KB base_model.cpython-36.pyc 8KB __pycache__ torch_utils.cpython-36.pyc 3KB data_generator.cpython-36.pyc 8KB torch_utils.cpython-37.pyc 3KB __init__.cpython-37.pyc 121B __init__.cpython-36.pyc 110B torch_utils.py 4KB requirements.txt 58B .gitignore 1KB README.md 3KB资源评论
