Merge d0f4b58 into 7209b99

.travis.yml

@@ -4,6 +4,9 @@
 language: python
 python:
   - 3.6
+  - 3.7
+  - 3.8
+  - 3.9
 
 dist: xenial
 services:

README.md

@@ -1,4 +1,5 @@
 # Py-FEAT  
+[![Package versioning](https://img.shields.io/pypi/v/py-feat.svg)](https://pypi.org/project/py-feat/)
 [![Build Status](https://api.travis-ci.org/cosanlab/feat.svg?branch=master)](https://travis-ci.org/cosanlab/feat/)
 [![Coverage Status](https://coveralls.io/repos/github/cosanlab/py-feat/badge.svg?branch=master)](https://coveralls.io/github/cosanlab/py-feat?branch=master)
 Python Facial Expression Analysis Toolbox (FEAT)

feat/detector.py

@@ -634,7 +634,7 @@ def detect_video(self, inputFname, outputFname=None, skip_frames=1, verbose=Fals
                 emotion_columns=FEAT_EMOTION_COLUMNS,
                 facebox_columns=FEAT_FACEBOX_COLUMNS,
                 landmark_columns=openface_2d_landmark_columns,
-                time_columns=FACET_TIME_COLUMNS,
+                time_columns=FEAT_TIME_COLUMNS,
                 detector="Feat",
             )
 
@@ -688,23 +688,3 @@ def detect_image(self, inputFname, outputFname=None, verbose=False):
                 time_columns=FACET_TIME_COLUMNS,
                 detector="Feat",
             )
-# %%
-# Test case:
-if __name__ == '__main__':
-    A01 = Detector(face_model='RetinaFace',emotion_model='svm', landmark_model="MobileFaceNet", au_model='rf')
-    test_img = cv2.imread(r"F:\test_case\JinHyunCheong.jpg")
-    im01 = Image.open(r"F:\test_case\JinHyunCheong.jpg")
-    detected_faces = A01.detect_faces(frame=test_img)
-    landmarks = A01.detect_landmarks(
-                    frame=test_img, detected_faces=[detected_faces[0][0:4]]
-                )
-    ress = A01.detect_image(r"F:\test_case\JinHyunCheong.jpg")
-    print(ress)
-    #convex_hull, new_lands = A01.extract_face(frame=test_img, detected_faces=detected_faces[0], landmarks=landmarks, size_output=112)
-    #bk01,bk02 = A01.extract_hog(frame=convex_hull,visualize=True)
-    #A02 = LogisticClassifier()
-    #probbs = A02.detect_au(bk01, new_lands)
-    #print(probbs)
-    print("yes")
-    ress.plot_detections();
-# %%

feat/tests/test_detector.py

@@ -12,6 +12,7 @@
 import os
 import cv2
 import numpy as np
+import pytest
 
 inputFname = os.path.join(get_test_data_path(), "input.jpg")
 img01 = cv2.imread(inputFname)
@@ -211,6 +212,28 @@ def test_resmasknet():
     out = detector1.detect_image(inputFname)
     assert out.emotions()["happiness"].values > 0.5
 
+def test_emotionsvm():
+    inputFname = os.path.join(get_test_data_path(), "input.jpg")
+    detector1 = Detector(emotion_model="svm")
+    out = detector1.detect_image(inputFname)
+    assert out.emotions()["happiness"].values > 0.5
+
+def test_emotionrf():
+    # Emotion RF models is not good 
+    inputFname = os.path.join(get_test_data_path(), "input.jpg")
+    detector1 = Detector(emotion_model="rf")
+    out = detector1.detect_image(inputFname)
+    assert out.emotions()["happiness"].values > 0.0
+
+def test_wrongmodelname():
+    with pytest.raises(KeyError):
+        detector1 = Detector(emotion_model="badmodelname")
+
+def test_nofile():
+    with pytest.raises(FileNotFoundError):
+        inputFname = os.path.join(get_test_data_path(), "nosuchfile.jpg")
+        detector1 = Detector(emotion_model="svm")
+        out = detector1.detect_image(inputFname)
 
 def test_detect_image():
     # Test detect image

feat/tests/test_utils.py

@@ -10,6 +10,7 @@
 from os.path import join, exists
 from .utils import get_test_data_path
 from feat.utils import (
+    read_feat,
     read_facet,
     read_openface,
     read_affectiva,
@@ -18,10 +19,15 @@
     softmax,
     load_h5,
 )
+from feat import Fex
 from nltools.data import Adjacency
 import unittest
 
 
+def test_read_feat():
+    fex = read_feat(join(get_test_data_path(), "output.csv"))
+    assert type(fex) == Fex
+
 def test_utils():
     sample = read_openface(join(get_test_data_path(), "OpenFace_Test.csv"))
     lm_cols = ["x_" + str(i) for i in range(0, 68)] + [

feat/utils.py

@@ -241,10 +241,11 @@ def read_feat(fexfile):
         Fex of processed facial expressions
     """
     d = pd.read_csv(fexfile)
+    au_columns = [col for col in d.columns if "AU" in col]
     return feat.Fex(
         d,
         filename=fexfile,
-        au_columns=jaanet_AU_presence,
+        au_columns=au_columns,
         emotion_columns=FEAT_EMOTION_COLUMNS,
         landmark_columns=openface_2d_landmark_columns,
         facebox_columns=FEAT_FACEBOX_COLUMNS,

notebooks/_build/html/_sources/content/intro.md

@@ -1,11 +1,11 @@
 Py-Feat: Python Facial Expression Analysis Toolbox
 ============================
-
-[![Build Status](https://api.travis-ci.org/cosanlab/feat.svg?branch=master)](https://travis-ci.org/cosanlab/feat/)
+[![Package versioning](https://img.shields.io/pypi/v/py-feat.svg)](https://pypi.org/project/py-feat/)
+[![Build Status](https://api.travis-ci.org/cosanlab/py-feat.svg?branch=master)](https://travis-ci.org/cosanlab/py-feat/)
 [![Coverage Status](https://coveralls.io/repos/github/cosanlab/py-feat/badge.svg?branch=master)](https://coveralls.io/github/cosanlab/py-feat?branch=master)
-[![GitHub forks](https://img.shields.io/github/forks/cosanlab/py-feat)](https://github.com/cosanlab/feat/network)
-[![GitHub stars](https://img.shields.io/github/stars/cosanlab/py-feat)](https://github.com/cosanlab/feat/stargazers)
-[![GitHub license](https://img.shields.io/github/license/cosanlab/py-feat)](https://github.com/cosanlab/feat/blob/master/LICENSE)
+[![GitHub forks](https://img.shields.io/github/forks/cosanlab/py-feat)](https://github.com/cosanlab/py-feat/network)
+[![GitHub stars](https://img.shields.io/github/stars/cosanlab/py-feat)](https://github.com/cosanlab/py-feat/stargazers)
+[![GitHub license](https://img.shields.io/github/license/cosanlab/py-feat)](https://github.com/cosanlab/py-feat/blob/master/LICENSE)
 
 Py-Feat provides a comprehensive set of tools and models to easily detect facial expressions (Action Units, emotions, facial landmarks) from images and videos, preprocess & analyze facial expression data, and visualize facial expression data. 
 

notebooks/_build/html/content/analysis.html

@@ -369,8 +369,8 @@ <h1 class="site-logo" id="site-title">Py-Feat</h1>
     </a>
    </li>
    <li class="toc-h2 nav-item toc-entry">
-    <a class="reference internal nav-link" href="#intersubject-correlations">
-     Intersubject correlations
+    <a class="reference internal nav-link" href="#intersubject-or-intervideo-correlations">
+     Intersubject (or intervideo) correlations
     </a>
    </li>
   </ul>
@@ -894,6 +894,11 @@ <h2>Simple t-test<a class="headerlink" href="#simple-t-test" title="Permalink to
       <td>-49.805912</td>
       <td>2.660164e-12</td>
     </tr>
+    <tr>
+      <th>mean_AU05</th>
+      <td>-11.942339</td>
+      <td>8.022300e-07</td>
+    </tr>
     <tr>
       <th>mean_AU06</th>
       <td>3.905366</td>
@@ -904,11 +909,21 @@ <h2>Simple t-test<a class="headerlink" href="#simple-t-test" title="Permalink to
       <td>5.487334</td>
       <td>3.864728e-04</td>
     </tr>
+    <tr>
+      <th>mean_AU09</th>
+      <td>-22.647961</td>
+      <td>3.024835e-09</td>
+    </tr>
     <tr>
       <th>mean_AU10</th>
       <td>10.361594</td>
       <td>2.660301e-06</td>
     </tr>
+    <tr>
+      <th>mean_AU11</th>
+      <td>-42.691265</td>
+      <td>1.059418e-11</td>
+    </tr>
     <tr>
       <th>mean_AU12</th>
       <td>11.949046</td>
@@ -929,6 +944,11 @@ <h2>Simple t-test<a class="headerlink" href="#simple-t-test" title="Permalink to
       <td>-0.460219</td>
       <td>6.562759e-01</td>
     </tr>
+    <tr>
+      <th>mean_AU20</th>
+      <td>-45.470892</td>
+      <td>6.019391e-12</td>
+    </tr>
     <tr>
       <th>mean_AU23</th>
       <td>-0.300423</td>
@@ -939,6 +959,26 @@ <h2>Simple t-test<a class="headerlink" href="#simple-t-test" title="Permalink to
       <td>-4.533345</td>
       <td>1.419377e-03</td>
     </tr>
+    <tr>
+      <th>mean_AU25</th>
+      <td>3.119502</td>
+      <td>1.232852e-02</td>
+    </tr>
+    <tr>
+      <th>mean_AU26</th>
+      <td>-5.015769</td>
+      <td>7.232461e-04</td>
+    </tr>
+    <tr>
+      <th>mean_AU28</th>
+      <td>-7.169935</td>
+      <td>5.251451e-05</td>
+    </tr>
+    <tr>
+      <th>mean_AU43</th>
+      <td>-78.139016</td>
+      <td>4.660607e-14</td>
+    </tr>
   </tbody>
 </table>
 </div></div></div>
@@ -1076,15 +1116,23 @@ <h2>Regression<a class="headerlink" href="#regression" title="Permalink to this
       <th>AU01</th>
       <th>AU02</th>
       <th>AU04</th>
+      <th>AU05</th>
       <th>AU06</th>
       <th>AU07</th>
+      <th>AU09</th>
       <th>AU10</th>
+      <th>AU11</th>
       <th>AU12</th>
       <th>AU14</th>
       <th>AU15</th>
       <th>AU17</th>
+      <th>AU20</th>
       <th>AU23</th>
       <th>AU24</th>
+      <th>AU25</th>
+      <th>AU26</th>
+      <th>AU28</th>
+      <th>AU43</th>
     </tr>
   </thead>
   <tbody>
@@ -1093,30 +1141,46 @@ <h2>Regression<a class="headerlink" href="#regression" title="Permalink to this
       <td>0.066</td>
       <td>0.052</td>
       <td>-0.065</td>
+      <td>-0.251</td>
       <td>0.497</td>
       <td>0.291</td>
+      <td>0.105</td>
       <td>0.408</td>
+      <td>-0.002</td>
       <td>0.569</td>
       <td>0.022</td>
       <td>0.111</td>
       <td>0.101</td>
+      <td>0.038</td>
       <td>-0.029</td>
       <td>-0.284</td>
+      <td>0.305</td>
+      <td>0.140</td>
+      <td>-0.016</td>
+      <td>0.013</td>
     </tr>
     <tr>
       <th>t-stats</th>
       <td>8.221</td>
       <td>8.219</td>
       <td>-17.153</td>
+      <td>-31.413</td>
       <td>43.783</td>
       <td>33.364</td>
+      <td>21.079</td>
       <td>35.625</td>
+      <td>-2.671</td>
       <td>46.713</td>
       <td>4.088</td>
       <td>17.515</td>
       <td>15.673</td>
+      <td>7.927</td>
       <td>-5.063</td>
       <td>-22.483</td>
+      <td>19.415</td>
+      <td>16.822</td>
+      <td>-2.293</td>
+      <td>2.422</td>
     </tr>
     <tr>
       <th>p-values</th>
@@ -1128,19 +1192,28 @@ <h2>Regression<a class="headerlink" href="#regression" title="Permalink to this
       <td>0.000</td>
       <td>0.000</td>
       <td>0.000</td>
+      <td>0.008</td>
       <td>0.000</td>
       <td>0.000</td>
       <td>0.000</td>
       <td>0.000</td>
+      <td>0.000</td>
+      <td>0.000</td>
+      <td>0.000</td>
+      <td>0.000</td>
+      <td>0.000</td>
+      <td>0.022</td>
+      <td>0.016</td>
     </tr>
   </tbody>
 </table>
 </div></div></div>
 </div>
 </div>
-<div class="section" id="intersubject-correlations">
-<h2>Intersubject correlations<a class="headerlink" href="#intersubject-correlations" title="Permalink to this headline">¶</a></h2>
-<p>To compare the similarity of signals over time between subjects or videos, you can use the <code class="docutils literal notranslate"><span class="pre">isc</span></code> method. You can get a sense of how much two signals, such as a certain action unit activity, correlates over time. For example, if you want to see how AU01 activations are similar across the videos, you can do the following which shows that the temporal profile of AU01 activations form two clusters between the goodNews and the badNews conditions.</p>
+<div class="section" id="intersubject-or-intervideo-correlations">
+<h2>Intersubject (or intervideo) correlations<a class="headerlink" href="#intersubject-or-intervideo-correlations" title="Permalink to this headline">¶</a></h2>
+<p>To compare the similarity of signals over time between subjects or videos, you can use the <code class="docutils literal notranslate"><span class="pre">isc</span></code> method. You can get a sense of how much two signals, such as a certain action unit activity, correlates over time.</p>
+<p>In this example, we are calculating the ISC over videos. We want to check how similar AU01 activations are across videos so our session is set to the <code class="docutils literal notranslate"><span class="pre">input</span></code> which is the video name. Executing the <code class="docutils literal notranslate"><span class="pre">isc</span></code> method shows that the temporal profile of AU01 activations form two clusters between the goodNews and the badNews conditions.</p>
 <div class="cell docutils container">
 <div class="cell_input docutils container">
 <div class="highlight-ipython3 notranslate"><div class="highlight"><pre><span></span><span class="n">fex</span><span class="o">.</span><span class="n">sessions</span> <span class="o">=</span> <span class="n">fex</span><span class="o">.</span><span class="n">input</span><span class="p">()</span>

notebooks/_build/html/content/contribute.html

@@ -130,11 +130,6 @@ <h1 class="site-logo" id="site-title">Py-Feat</h1>
    Analyzing FEX data
   </a>
  </li>
- <li class="toctree-l1">
-  <a class="reference internal" href="analysis.html#two-sample-independent-t-test">
-   Two sample independent t-test
-  </a>
- </li>
  <li class="toctree-l1">
   <a class="reference internal" href="plotting.html">
    Plotting examples

notebooks/_build/html/content/detector.html

@@ -131,11 +131,6 @@ <h1 class="site-logo" id="site-title">Py-Feat</h1>
    Analyzing FEX data
   </a>
  </li>
- <li class="toctree-l1">
-  <a class="reference internal" href="analysis.html#two-sample-independent-t-test">
-   Two sample independent t-test
-  </a>
- </li>
  <li class="toctree-l1">
   <a class="reference internal" href="plotting.html">
    Plotting examples