From 66db70a61605b0876f6203f58a1517ee5b17b9bf Mon Sep 17 00:00:00 2001
From: Arjun V <arjunvkp@gmail.com>
Date: Mon, 16 Jul 2018 17:40:01 +0530
Subject: [PATCH 1/8] Update requirements.txt

Adding flask as a requirement
---
 requirements.txt | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/requirements.txt b/requirements.txt
index ccbe920..1ffe0bb 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,4 @@
-ImageHash==3.4
-scikit-learn==0.18
-numpy==1.11.1
+ImageHash
+scikit-learn
+numpy
+flask

From 44fc9e7935748d5537924ad35b2384d691ba7221 Mon Sep 17 00:00:00 2001
From: Arjun V <arjunvkp@gmail.com>
Date: Mon, 16 Jul 2018 17:41:06 +0530
Subject: [PATCH 2/8] Update readme.md

Correcting run command
---
 readme.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/readme.md b/readme.md
index 44c850f..42e0933 100644
--- a/readme.md
+++ b/readme.md
@@ -4,6 +4,6 @@ this is an unrealistically simple example, but this clusters "brain expanding" m
 
 install deps: `pip install -r requirements.txt`
 
-to run: `python cluster.py`
+to run: `python2 server.py`
 
-went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)
\ No newline at end of file
+went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)

From f9084db5ecba6412cb0252621dbc243cbbb1a146 Mon Sep 17 00:00:00 2001
From: Arjun V <arjunv@live.in>
Date: Wed, 18 Jul 2018 11:54:46 +0530
Subject: [PATCH 3/8] Update readme.md

---
 readme.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/readme.md b/readme.md
index 42e0933..1a90887 100644
--- a/readme.md
+++ b/readme.md
@@ -4,6 +4,6 @@ this is an unrealistically simple example, but this clusters "brain expanding" m
 
 install deps: `pip install -r requirements.txt`
 
-to run: `python2 server.py`
+to run: `python3 server.py`
 
 went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)

From e2ae27e7c4f6ce6ce82e678e6f0b0bb9b3ae3962 Mon Sep 17 00:00:00 2001
From: arjun v <arjunvkp@gmail.com>
Date: Thu, 19 Jul 2018 12:49:10 +0530
Subject: [PATCH 4/8] Added clustering images to directory

Changed default action to move images into a directory named "clusters".
Accepts arguments for running the flask server.
---
 .gitignore            |  4 ++
 cluster.py            | 11 +++++-
 data/.gitignore       |  2 -
 readme.md             | 18 +++++++--
 requirements.txt      |  7 ++--
 server.py             | 88 +++++++++++++++++++++++++++++++------------
 static/img/.gitignore |  2 -
 7 files changed, 96 insertions(+), 36 deletions(-)
 create mode 100644 .gitignore
 delete mode 100644 data/.gitignore
 delete mode 100644 static/img/.gitignore

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..b5934e4
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,4 @@
+.directory
+data/
+static/img/
+clusters/
\ No newline at end of file
diff --git a/cluster.py b/cluster.py
index 9983a3b..a27a93d 100644
--- a/cluster.py
+++ b/cluster.py
@@ -1,3 +1,4 @@
+import os
 import imagehash
 import numpy as np
 from PIL import Image
@@ -38,4 +39,12 @@ def cluster(mat, fnames, eps, min_samples):
     clusters = defaultdict(list)
     for i, lbl in enumerate(labels):
         clusters[lbl].append(fnames[i])
-    return clusters
\ No newline at end of file
+    return clusters
+
+def create_directory(directory):
+    try:
+        os.makedirs(directory)
+        return True
+    except OSError:
+        if not os.path.isdir(directory): return False
+        return True
\ No newline at end of file
diff --git a/data/.gitignore b/data/.gitignore
deleted file mode 100644
index c96a04f..0000000
--- a/data/.gitignore
+++ /dev/null
@@ -1,2 +0,0 @@
-*
-!.gitignore
\ No newline at end of file
diff --git a/readme.md b/readme.md
index 44c850f..ed803d1 100644
--- a/readme.md
+++ b/readme.md
@@ -2,8 +2,20 @@ super simple example of image clustering with image hashes ([wavelet hashing](ht
 
 this is an unrealistically simple example, but this clusters "brain expanding" memes from "political compass" memes pretty well.
 
-install deps: `pip install -r requirements.txt`
+install deps: `pip3 install -r requirements.txt`
 
-to run: `python cluster.py`
+went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)
 
-went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)
\ No newline at end of file
+to run: `python3 server.py`
+
+```
+usage: server.py [-h] [-s] [--cluster distance] [--min-samples count]
+
+Image Cluster
+
+optional arguments:
+  -h, --help           show this help message and exit
+  -s, --server         run flask server (default: False)
+  --cluster distance   cluster distance (default: 1)
+  --min-samples count  minimum number of samples in a cluster (default: 1)
+```
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index ccbe920..955d3bb 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,4 @@
-ImageHash==3.4
-scikit-learn==0.18
-numpy==1.11.1
+ImageHash
+scikit-learn
+numpy
+flask
\ No newline at end of file
diff --git a/server.py b/server.py
index 6137ef3..bab3103 100644
--- a/server.py
+++ b/server.py
@@ -1,16 +1,38 @@
 import json
+import shutil
 import numpy as np
 from glob import glob
-from flask import Flask, request, render_template, abort
-from cluster import compute_hashes, compute_dists, cluster
+from cluster import compute_hashes, compute_dists, cluster, create_directory
+
+import argparse
+parser = argparse.ArgumentParser(description='Image Cluster')
+
+# begin arguments
+parser.add_argument('-s', '--server', action='store_true',
+                    help='run flask server (default: %(default)s)\n\n')
+parser.add_argument('--cluster', metavar='distance', default=1,
+                    type=float, help='cluster distance (default: %(default)s)')
+parser.add_argument('--min-samples', metavar='count', default=1,
+                    type=int, help='minimum number of samples in a cluster (default: %(default)s)')
+
+args = parser.parse_args()
 
 
 if __name__ == '__main__':
+
+    if not create_directory('static/img/'):
+        print('Couldn\'t create static images directory!')
+        raise SystemExit
+
     try:
         mat = np.load('data/dist_mat.npy')
         fnames = json.load(open('data/fnames.json', 'r'))
     except FileNotFoundError:
-        print('computing hashes & distance matrix...')
+        if not create_directory('data'):
+            print('Couldn\'t create data directory!')
+            raise SystemExit
+
+        print('Computing hashes & distance matrix...')
         hashes, fnames = compute_hashes(glob('static/img/*'))
         mat = compute_dists(hashes)
 
@@ -19,25 +41,41 @@
             json.dump(fnames, f)
         np.save('data/dist_mat.npy', mat)
 
-    clusters = {}
-    app = Flask(__name__)
-
-    @app.route('/', methods=['GET', 'POST'])
-    def index():
-        global clusters
-        if request.method == 'POST':
-            eps = float(request.form.get('eps', 20))
-            min_samples = int(request.form.get('min_samples', 2))
-            clusters = cluster(mat, fnames, eps, min_samples)
-            print(clusters)
-        return render_template('index.html', clusters=clusters)
-
-    @app.route('/cluster/<int:id>', methods=['GET', 'POST'])
-    def view_cluster(id):
-        try:
-            print(clusters[id])
-            return render_template('cluster.html', cluster=clusters[id])
-        except KeyError:
-            abort(404)
-
-    app.run(host='0.0.0.0', port=5001)
+
+    if not args.server:
+        eps = args.cluster
+        min_samples = args.min_samples
+        clusters = cluster(mat, fnames, eps, min_samples)
+
+        print('Generated {} clusters!'.format(len(clusters)))
+        print('Moving images to "clusters" directory..')
+        for cluster in clusters:
+            create_directory('clusters/{}'.format(cluster))
+            for image in clusters[cluster]:
+                shutil.move(image, 'clusters/{}/'.format(cluster))
+        print('Created {} clusters and stored the images in "clusters" directory!'.format(len(clusters)))
+
+    else:
+        from flask import Flask, request, render_template, abort
+        clusters = {}
+        app = Flask(__name__)
+
+        @app.route('/', methods=['GET', 'POST'])
+        def index():
+            global clusters
+            if request.method == 'POST':
+                eps = float(request.form.get('eps', 20))
+                min_samples = int(request.form.get('min_samples', 2))
+                clusters = cluster(mat, fnames, eps, min_samples)
+                print(clusters)
+            return render_template('index.html', clusters=clusters)
+
+        @app.route('/cluster/<int:id>', methods=['GET', 'POST'])
+        def view_cluster(id):
+            try:
+                print(clusters[id])
+                return render_template('cluster.html', cluster=clusters[id])
+            except KeyError:
+                abort(404)
+
+        app.run(host='0.0.0.0', port=5002)
diff --git a/static/img/.gitignore b/static/img/.gitignore
deleted file mode 100644
index c96a04f..0000000
--- a/static/img/.gitignore
+++ /dev/null
@@ -1,2 +0,0 @@
-*
-!.gitignore
\ No newline at end of file

From de3a965eb21012d0f0563bef00fdfee837a874ca Mon Sep 17 00:00:00 2001
From: arjun v <arjunvkp@gmail.com>
Date: Thu, 19 Jul 2018 13:00:18 +0530
Subject: [PATCH 5/8] Added relevant help in README.md

---
 readme.md => README.md | 10 +++++-----
 requirements.txt       |  2 +-
 tests.py               | 33 ---------------------------------
 3 files changed, 6 insertions(+), 39 deletions(-)
 rename readme.md => README.md (64%)
 delete mode 100644 tests.py

diff --git a/readme.md b/README.md
similarity index 64%
rename from readme.md
rename to README.md
index 9b08200..1b9ae52 100644
--- a/readme.md
+++ b/README.md
@@ -1,12 +1,12 @@
 super simple example of image clustering with image hashes ([wavelet hashing](https://fullstackml.com/wavelet-image-hash-in-python-3504fdd282b5) in particular) and DBSCAN.
 
-this is an unrealistically simple example, but this clusters "brain expanding" memes from "political compass" memes pretty well.
-
 install deps: `pip3 install -r requirements.txt`
 
-went with wavelet hashing because of the results of my unscientific experimentation (see `tests.py`)
-
-to run: `python3 server.py`
+to run:
+ - store all images in `static/img/` directory
+ - empty `data` directory if u wish to do a re-run of clustering
+ - run `python3 server.py --server` to do an interactive run and decide the cluster distance and min-samples count.
+ - run without `--server` argument to move the images to `clusters` directory
 
 ```
 usage: server.py [-h] [-s] [--cluster distance] [--min-samples count]
diff --git a/requirements.txt b/requirements.txt
index 1ffe0bb..955d3bb 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,4 @@
 ImageHash
 scikit-learn
 numpy
-flask
+flask
\ No newline at end of file
diff --git a/tests.py b/tests.py
deleted file mode 100644
index a845b83..0000000
--- a/tests.py
+++ /dev/null
@@ -1,33 +0,0 @@
-import imagehash
-from PIL import Image
-from glob import glob
-from itertools import combinations
-
-imgs = [(Image.open(fname), fname) for fname in glob('example/*.jpg')]
-
-for name, hashfunc in [
-    ('phash', imagehash.phash),
-    ('ahash', imagehash.average_hash),
-    ('dhash', imagehash.dhash),
-    ('whash', imagehash.whash),
-    ('whash-db4', lambda img: imagehash.whash(img, mode='db4'))
-]:
-    print('hashfunc:', name)
-    hashes = [(hashfunc(im), fname) for im, fname in imgs]
-    pairs = combinations(hashes, 2)
-    true_sim_dists = []
-    true_dif_dists = []
-    for (a, a_n), (b, b_n) in pairs:
-        # pc=political compass
-        if 'pc' in a_n or 'pc' in b_n:
-            true_dif_dists.append(a - b)
-        else:
-            true_sim_dists.append(a - b)
-        print(a_n, b_n, ':', a - b)
-
-    # want the `max sim dist` to be less than
-    # the `min dif dist`. the bigger the gap, the better
-    print('min dif dist:', min(true_dif_dists))
-    print('max sim dist:', max(true_sim_dists))
-    print('clean split?:', max(true_sim_dists) < min(true_dif_dists))
-    print('---')
\ No newline at end of file

From 06c69d9b327be2fdf9e416dcaf36889bea9f791a Mon Sep 17 00:00:00 2001
From: arjun v <arjunvkp@gmail.com>
Date: Thu, 19 Jul 2018 13:05:58 +0530
Subject: [PATCH 6/8] Adding subdirectories

---
 .gitignore | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/.gitignore b/.gitignore
index b5934e4..08131f9 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,4 +1,8 @@
 .directory
-data/
-static/img/
-clusters/
\ No newline at end of file
+data/*
+static/img/*
+clusters/*
+
+!static/img
+!data
+!clusters
\ No newline at end of file

From 7a06698f3e999380bea5a92bb487110f3aafee11 Mon Sep 17 00:00:00 2001
From: arjun v <arjunvkp@gmail.com>
Date: Thu, 19 Jul 2018 13:10:21 +0530
Subject: [PATCH 7/8] Adding subdirectories

---
 .gitignore            | 9 ++-------
 static/img/.gitignore | 2 ++
 2 files changed, 4 insertions(+), 7 deletions(-)
 create mode 100644 static/img/.gitignore

diff --git a/.gitignore b/.gitignore
index 08131f9..51e826a 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,8 +1,3 @@
 .directory
-data/*
-static/img/*
-clusters/*
-
-!static/img
-!data
-!clusters
\ No newline at end of file
+*.pyc
+__pycache__/
\ No newline at end of file
diff --git a/static/img/.gitignore b/static/img/.gitignore
new file mode 100644
index 0000000..c96a04f
--- /dev/null
+++ b/static/img/.gitignore
@@ -0,0 +1,2 @@
+*
+!.gitignore
\ No newline at end of file

From 87071f6e142ab04ab8586245d0d8a1691290a7e0 Mon Sep 17 00:00:00 2001
From: arjun v <arjunvkp@gmail.com>
Date: Thu, 19 Jul 2018 13:11:13 +0530
Subject: [PATCH 8/8] Adding subdirectories

---
 .gitignore | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/.gitignore b/.gitignore
index 51e826a..f0d571f 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,5 @@
 .directory
 *.pyc
-__pycache__/
\ No newline at end of file
+__pycache__/
+data/
+clusters/
\ No newline at end of file