You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
- Add no_duplicates page from #71
- Improve fuzzy search for the new logic of multiple hashes
- Add table layout of dupliacates
- Extend documentation
- pylint and pycodestyle fixed, add ubuntu 22.04 test scripts
- Adapt tests
This Python script finds duplicate images using a [perspective hash (pHash)](http://www.phash.org) to compare images. pHash ignores the image size and file size and instead creates a hash based on the pixels of the image. This allows you to find duplicate pictures that have been rotated, have changed metadata, and slightly edited.
5
+
This Python script finds duplicate files:
6
+
- any file by exact match of blake2b hash
7
+
- images using a [perspective hash (pHash)](http://www.phash.org) to compare images. pHash ignores the image size and file size and instead creates a hash based on the pixels of the image. This allows you to find duplicate pictures that have been rotated, have changed metadata, and slightly edited.
8
+
- videos by extracting N frames at respective times and hashing them with perspective hash (see above)
6
9
7
-
This script hashes images added to it, storing the hash into a database (MongoDB). To find duplicate images, hashes are compared. If the hash is the same between two images, then they are marked as duplicates. A web interface is provided to delete duplicate images easily. If you are feeling lucky, there is an option to automatically delete duplicate files.
10
+
This script hashes files added to it, storing the hashes into a database (MongoDB). To find duplicate files, hashes are compared. If the hash is the same between two files, then they are marked as duplicates. A web interface is provided to delete duplicate files easily. If you are feeling lucky, there is an option to automatically delete duplicate files.
8
11
9
-
As a word of caution, pHash is not perfect. I have found that duplicate pictures sometimes have different hashes and similar (but not the same) pictures have the same hash. This script is a great starting point for cleaning your photo library of duplicate pictures, but make sure you look at the pictures before you delete them. You have been warned! I hold no responsibility for any family memories that might be lost because of this script.
10
-
11
-
This script has only been tested with Python 3 and is still pretty rough around the edges. Use at your own risk.
12
+
As a word of caution, pHash is not perfect. I have found that duplicate pictures sometimes have different hashes and similar (but not the same) pictures have the same hash. This script is a great starting point for cleaning your photo or video library of duplicate pictures, but make sure you look at the pictures before you delete them. You have been warned! I hold no responsibility for any family memories that might be lost because of this script.
12
13
13
14
## Requirements
14
15
15
-
This script requires MongoDB, Python 3.4 or higher, and a few Python modules, as found in `requirements.txt`.
16
-
16
+
This script requires MongoDB, Python 3.12 or higher, and a few Python modules, as found in `requirements.txt`.
17
17
18
18
## Quick Start
19
19
20
20
I suggest you read the usage, but here are the steps to get started right away. These steps assume that MongoDB is already installed on the system.
21
21
22
-
First, install this script. This can be done by either cloning the repository or [downloading the script](https://github.com/philipbl/duplicate-images/archive/master.zip).
22
+
First, install this script. This can be done by either cloning the repository or [downloading the script](https://github.com/bolshevik/duplicate-images/archive/master.zip).
Next, download all required modules. This script has only been tested with Python 3. I would suggest that you make a virtual environment, setting Python 3 as the default python executable (`mkvirtualenv --python=/usr/local/bin/python3 <name>`)
@@ -34,26 +34,6 @@ Last, run script:
34
34
python duplicate_finder.py
35
35
```
36
36
37
-
## On Ubuntu 18.04
38
-
39
-
```bash
40
-
# Install Mongo and pip
41
-
sudo apt -y install mongodb-server python3-pip
42
-
# Disable Mongo service autostart
43
-
sudo systemctl disable mongodb.service
44
-
# Stop Mongo service
45
-
sudo service mongodb stop
46
-
```
47
-
48
-
Python 2 is the default version of Python, so we have to call `python3` explicitely:
49
-
50
-
```bash
51
-
# Install dependencies with Python 3
52
-
pip3 install -r requirements.txt
53
-
# “python duplicate_finder.py” will fail, so we have to use Python 3 for every call:
--db=<db_path> The location of the database or a MongoDB URI. (default: ./db)
86
67
87
-
--parallel=<num_processes> The number of parallel processes to run to hash the image
88
-
files (default: number of CPUs).
68
+
--parallel=<num_processes> The number of parallel processes to run to hash the files
69
+
(default: number of CPUs).
89
70
90
71
find:
91
-
--threshold=<num>Image matching threshold. Number of different bits in Hamming distance. False positives are possible.
72
+
--threshold=<num>Hash matching threshold. Number of different bits in Hamming distance. False positives are possible.
92
73
--print Only print duplicate files rather than displaying HTML file
93
-
--delete Move all found duplicate pictures to the trash. This option takes priority over --print.
74
+
--delete Move all found duplicate files to the trash. This option takes priority over --print.
94
75
--match-time Adds the extra constraint that duplicate images must have the
95
76
same capture timesin order to be considered.
96
77
--trash=<trash_path> Where files will be put when they are deleted (default: ./Trash)
@@ -101,14 +82,14 @@ Options:
101
82
python duplicate_finder.py add /path/to/images
102
83
```
103
84
104
-
When a path is added, image files are recursively searched for. In particular, `JPEG`, `PNG`, `GIF`, and `TIFF`images are searched for. Any image files found will be hashed. Adding a path uses 8 processes (by default) to hash images in parallel so the CPU usage is very high.
85
+
When a path is added, files are recursively searched for. Binary content hash is applied to all files, for image files like `JPEG`, `PNG`, `GIF`, and `TIFF`the perceptive hash is applied. Video hash is applied to video files. Adding a path uses 8 processes (by default) to hash images in parallel so the CPU usage is very high.
105
86
106
87
### Remove
107
88
```bash
108
89
python duplicate_finder.py remove /path/to/images
109
90
```
110
91
111
-
A path can be removed from the database. Any image inside that path will be removed from the database.
92
+
A path can be removed from the database. Any file inside that path will be removed from the database.
Finds duplicate pictures that have been hashed. This will find images that have the same hash stored in the database. There are a few options associated with `find`. By default, when this command is run, a webpage is displayed showing duplicate pictures and a server is started that allows for the pictures to be deleted (images are not actually deleted, but moved to a trash folder -- I really don't want you to make a mistake). The first option, **`--print`**, prints all duplicate pictures and does not display a webpage or start the server. **`--delete`** automatically moves all duplicate images found to the trash. Be careful with this one. **`--match-time`** adds the extra constraint that images must have the same EXIF time stamp to be considered duplicate pictures. Last, `--trash=<trash_path>` lets you select a path to where you want files to be put when they are deleted. The default trash location is `./Trash`.
120
+
Finds duplicate files that have been hashed. This will find files that have the same hash stored in the database. There are a few options associated with `find`. By default, when this command is run, a webpage is displayed showing duplicate files and a server is started that allows for the files to be deleted (files are not actually deleted, but moved to a trash folder -- I really don't want you to make a mistake). The first option, **`--print`**, prints all duplicate files and does not display a webpage or start the server. **`--delete`** automatically moves all duplicate files found to the trash. Be careful with this one. **`--match-time`** adds the extra constraint that images must have the same EXIF time stamp to be considered duplicate pictures. `--trash=<trash_path>` lets you select a path to where you want files to be put when they are deleted. The default trash location is `./Trash`. Last, `--threshold=<num>` specifies number of bits of Hamming distance to run fuzzy matching of hashes.
0 commit comments