Implement logging system and support debug levels, closes #6

actions.mk

@@ -63,14 +63,18 @@ graph_reconstruction:
 		--shape_type imm \
 		--n_components 6
 
+.PHONY:= test
 test:
 	python -m py.test -f src/test/*_test.py
 
+.PHONY:= server
 server:
 	(cd src/; python -m tornado.autoreload server.py)
 
+.PHONY:= ember
 ember:
 	(cd viewer; ember server);
 
+.PHONY:= ctags
 ctags:
 	ctags -R --python-kinds=-i src

src/aam.py

 """
 .. module:: active_appearance_model
-   :platform: Unix, Windows
+   :platform: Unix
    :synopsis: Contains the aam data format abstraction layer
 
 """
 
-import logging
 import numpy as np
 from matplotlib.tri import Triangulation
 import cv2
@@ -13,13 +12,7 @@ import cv2
 # local imports
 import pca
 import reconstruction.texture as tx
-
-
-logging.basicConfig(
-    level=logging.INFO,
-    format='%(asctime)s %(levelname)s %(name)s: %(message)s'
-)
-logger = logging.getLogger(__name__)
+from settings import logger
 
 
 class AAMPoints():

src/main.py

 #!/usr/local/bin/python
 # python std
 import argparse
-import logging
 import importlib
 
 # installed packages
@@ -14,9 +13,7 @@ import aam
 
 from reconstruction import reconstruction
 
-logging.basicConfig(level=logging.INFO,
-                    format='%(asctime)s %(levelname)s %(name)s: %(message)s')
-logger = logging.getLogger(__name__)
+from settings import logger
 
 
 def add_parser_options():
@@ -104,7 +101,7 @@ def save_pca_model_texture(args):
     assert args.shape_type, '--shape_type the type of dataset, see datasets module'
 
     dataset_module = import_dataset_module(args.shape_type)
-    shape_model = pca.PcaModel(args.model_shape_file)
+    shape_model = pca.PCAModel(args.model_shape_file)
     mean_points = dataset_module.IMMPoints(points_list=shape_model.mean_values)
 
     textures = aam.build_texture_feature_vectors(
@@ -174,8 +171,8 @@ def generate_call_graph(args):
     graphviz = GraphvizOutput(output_file='filter_none.png')
 
     with PyCallGraph(output=graphviz):
-        shape_model = pca.PcaModel(args.model_shape_file)
-        texture_model = pca.PcaModel(args.model_texture_file)
+        shape_model = pca.PCAModel(args.model_shape_file)
+        texture_model = pca.PCAModel(args.model_texture_file)
 
         input_points = dataset_module.IMMPoints(filename='data/imm_face_db/40-3m.asf')
         input_image = input_points.get_image()
@@ -198,8 +195,8 @@ def show_reconstruction(args):
 
     dataset_module = import_dataset_module(args.shape_type)
 
-    shape_model = pca.PcaModel(args.model_shape_file)
-    texture_model = pca.PcaModel(args.model_texture_file)
+    shape_model = pca.PCAModel(args.model_shape_file)
+    texture_model = pca.PCAModel(args.model_texture_file)
 
     input_points = dataset_module.IMMPoints(
         filename='data/imm_face_db/40-3m.asf'

src/pca.py

 import numpy as np
 
+from settings import logger
 
-class PcaModel:
+
+class PCAModel:
     """
     Abstraction for a pca model file. The pca model is stored in a numpy file
     using numpy.save. The following information is stored:
@@ -15,7 +17,7 @@ class PcaModel:
 
 
     Examples:
-        pca = PcaModel(path_to_numpy_model_file)
+        pca = PCAModel(path_to_numpy_model_file)
     """
     def __init__(self, filename=None):
         self.filename = filename
@@ -51,8 +53,7 @@ class PcaModel:
         assert hasattr(self, 'mean_values')
         assert hasattr(self, 'triangles')
 
-        saving = np.asarray(
-            [
+        saving = np.asarray([
                 self.Vt,
                 self.s,
                 self.n_components,
@@ -65,23 +66,23 @@ class PcaModel:
 
     def load(self):
         """
-        Loads the numpy file, see PcaModel whichs uses this function to load
+        Loads the numpy file, see PCAModel whichs uses this function to load
         the PCA Model data.
 
         Returns:
             (tuple): Vt, s, n_components, mean_values and triangles
             Vt (numpy ndarray): Two dimensional array with dimensions
             (n_features, n_features)
-            n_components: number of components needed to cover .90 percent of the
-            variance
-            mean_values (numpy ndarray): mean values of the features of the model,
-            this should have dimensions (n_featurs, )
+            n_components: number of components needed to cover .90 percent of
+            the variance
+            mean_values (numpy ndarray): mean values of the features of the
+            model, this should have dimensions (n_features, )
             triangles: a list of lists of indices that form a triangles in the
             AAM list.
 
         Examples:
             We would advise not to use this function directly but to use the
-            PcaModel. See the :class:`PcaModel`
+            PCAModel. See the :class:`PCAModel`
 
         """
         pca_model = np.load(self.filename)
@@ -94,12 +95,29 @@ class PcaModel:
 
 def pca(data, mean_values, variance_percentage=90):
     """
-    Perform Singlar Value Decomposition
+    Perform Singlar Value Decomposition which we see as a PCA analysis
+    We calculate how many components are needed to get `variance_percentage`
+    (default is 90 percent).
+
+    Args:
+        data(ndarray): list of flattened feature vectors.
+        mean_values(ndarray): mean of all data flattened feature vectors,
+        in the same order.
+
+    Kwargs:
+        variance_percentage(int): is to calculate how many components you would
+        need to keep 90 (default is 90) percent of the variance. Note that we
+        do not alter any data, just return extra information in the form of
+        `n_components`, so that the user knows how much components it could
+        keep or to discard to still have 90 percent of the variance.
 
     Returns:
+        tuple of:
             U (ndarray): U matrix
-        s (ndarray): 1d singular values (diagonal in array form)
+            s (ndarray): 1d singular values in flattened form.
             Vt (ndarray): Vt matrix
+            n_components(int): The amount of components that (together) form
+            `variance_percentage` of variance.
     """
     # subtract mean
     zero_mean = data - mean_values
@@ -115,6 +133,7 @@ def pca(data, mean_values, variance_percentage=90):
         i += 1
 
     n_components = i
+    logger.debug('%s components form %s of the variance', n_components, variance_percentage)
 
     return U, s, Vt, n_components
 
@@ -124,13 +143,13 @@ def reconstruct(feature_vector, Vt, mean_values, n_components=None):
     Reconstruct with U, s, Vt
 
     Args:
-        U (numpy ndarray): One feature vector from the reduced SVD.
+        U (numpy ndarray): One feature vector from the SVD.
             U should have shape (n_features,), (i.e., one dimensional)
         s (numpy ndarray): The singular values as a one dimensional array
         Vt (numpy ndarray): Two dimensional array with dimensions
         (n_features, n_features)
-        mean_values (numpy ndarray): mean values of the features of the model,
-        this should have dimensions (n_features, )
+        mean_values (numpy ndarray): mean values of the features of the
+        model, this should have dimensions (n_features, )
 
     """
 

src/reconstruct.py

-def reconstruct(model_shape_file, model_texture_file, image, asf_file):
-    #assert args.model_shape_file, '--model_texture_file needs to be provided to save the pca model'
-    #assert args.model_texture_file, '--model_texture_file needs to be provided to save the pca model'
-
-    Vt_shape, s, n_shape_components, mean_value_points, triangles = pca.load(args.model_shape_file)
-    Vt_texture, s_texture, n_texture_components, mean_values_texture, _ = pca.load(args.model_texture_file)
-
-    InputPoints = imm.IMMPoints(filename=asf_file)
-    input_image = InputPoints.get_image()
-
-    MeanPoints = imm.IMMPoints(points_list=mean_value_points)
-    MeanPoints.get_scaled_points(input_image.shape)
-
-    while True:
-        utils.reconstruct_texture(
-            input_image,  # src image
-            input_image,  # dst image
-            Vt_texture,   # Vt
-            InputPoints,  # shape points input
-            MeanPoints,   # shape points mean
-            mean_values_texture,  # mean texture
-            triangles,  # triangles
-            n_texture_components  # learned n_texture_components
-        )
-
-        dst = utils.get_texture(MeanPoints, mean_values_texture)
-
-        cv2.imshow('original', InputPoints.get_image())
-        cv2.imshow('reconstructed', input_image)
-        cv2.imshow('main face', dst)
-
-        k = cv2.waitKey(0) & 0xFF
-
-        if k == 27:
-            break
-
-    cv2.destroyAllWindows()

src/server.py

 import json
 import os.path
 import base64
-from cStringIO import StringIO
 from glob import glob
 
 import cv2
@@ -31,8 +30,8 @@ class ImageWebSocketHandler(websocket.WebSocketHandler):
         model_texture_file = '{}/pca_texture_model.npy'.format(FILES_DIR)
         model_shape_file = '{}/pca_shape_model.npy'.format(FILES_DIR)
 
-        self.shape_model = pca.PcaModel(model_shape_file)
-        self.texture_model = pca.PcaModel(model_texture_file)
+        self.shape_model = pca.PCAModel(model_shape_file)
+        self.texture_model = pca.PCAModel(model_texture_file)
 
         websocket.WebSocketHandler.__init__(self, *args, **kwargs)
 

src/settings.py

+"""
+.. module:: settings_module
+   :platform: Unix
+   :synopsis: This module contains global settings.
+
+"""
+import logging
+import logging.config
+import os
+
+#logging.basicConfig(level=logging.INFO,
+#                    format='%(asctime)s %(levelname)s %(name)s: %(message)s')
+
+logging.config.dictConfig({
+    'version': 1,
+    'disable_existing_loggers': False,  # this fixes the problem
+
+    'formatters': {
+        'standard': {
+            'format': '%(asctime)s %(levelname)s %(module)s: %(message)s'
+        },
+    },
+    'handlers': {
+        'default': {
+            'level': logging.INFO,
+            'formatter': 'standard',
+            'class': 'logging.StreamHandler',
+        },
+        'debug': {
+            'level': logging.DEBUG,
+            'formatter': 'standard',
+            'class': 'logging.StreamHandler',
+        },
+    },
+    'loggers': {
+        'root': {
+            'handlers': ['default'],
+            'level': logging.INFO,
+            'propagate': True
+        },
+        'debug': {
+            'handlers': ['debug'],
+            'level': logging.DEBUG,
+            'propagate': False
+        }
+    }
+})
+
+logger = logging.getLogger('root')
+#logger.setLevel(logging.DEBUG)
+
+if os.environ.get('DEBUG', False):
+    logger = logging.getLogger('debug')

src/test/aam_test.py

@@ -48,8 +48,8 @@ def test_zero_mean_aan():
 
 
 def test_build_texture_feature_vectors():
-    shape_model = pca.PcaModel('data/test_data/pca_shape_model.npy')
-    texture_model = pca.PcaModel('data/test_data/pca_texture_model.npy')
+    shape_model = pca.PCAModel('data/test_data/pca_shape_model.npy')
+    texture_model = pca.PCAModel('data/test_data/pca_texture_model.npy')
 
     input_points = imm.IMMPoints(filename='data/imm_face_db/40-3m.asf')
     input_image = input_points.get_image()