Metadata-Version: 2.1
Name: ddup
Version: 0.0.2
Summary: UNKNOWN
Home-page: UNKNOWN
License: UNKNOWN
Description: [toc]
        # dedup
        
        Compare images in two image lists and find dulplicate ones. <br/>
        
        ## Install
        
        - Clone the repo
        - Install with
        ```bash
        python setup.py install
        ```
        
        ## Usage
        
        - Use it in command line
        ```bash
        dedup path1 path2 save_dir [--log]
        ```
        Compare result will be sved in `dedup.json` in the specified `save_dir`
        
        ## Example
        
        ### Input
        - `path1` <br/>
        imglist1.txt
        ```
        /mnt/Storage/test1/000001.jpg
        /mnt/Storage/test1/000002.jpg
        ```
        - `path2` <br/>
        imglist2.txt
        ```
        /mnt/Storage/test2/000001.jpg
        /mnt/Storage/test2/000002.jpg
        /mnt/Storage/test2/000003.jpg
        ```
        - `save_dir`
        ```
        /mnt/Storage
        ```
        
        ### Output
        - `imglist1_phash.hdf5` <br/>
        Store the phashes of images in imglist1 in .hdf5 format. <br/>
        - `imglist2_phash.hdf5` <br/>
        Store the phashes of images in imglist2 in .hdf5 format. <br/>
        - `dedup.json` <br/>
        Store the comparision results in json format. Each image in list1 will correspond to one or more images in list2 if they are similar.
        ```json
        {
            "/mnt/Storage/test1/000001.jpg":[
                "/mnt/Storage/test2/000001.jpg"
            ],
            "/mnt/Storage/test1/000002.jpg":[
                "/mnt/Storage/test2/000002.jpg"
            ]
        }
        ```
        
        ## Parameters
        - `path1` <br/>
        Path of the first image list. <br/>
        Image paths in the list should be **Absolute path**. <br/>
        - `path2` <br/>
        Path of the second image list. <br/>
        Image paths in the list should be **Absolute path**. <br/>
        - `save_dir` <br/>
        Folder path to save the results. <br/>
        - `--log`[optional] <br/>
        With this option added, the program will print detail log for each thread and each dulplicate image pair.
        This may cause message flush on screen so it is recommended to pipe it in to a log file.
Platform: UNKNOWN
Description-Content-Type: text/markdown
