dict handling down to and from analysis routines (#19)

.flake8

@@ -1,6 +1,5 @@
 [flake8]
-ignore = E203, E266, E402, E501, W503, F403, F401, F841
+ignore = F401, E402, E501
 exclude = .git,__pycache__,.ipynb_checkpoints
 max-line-length = 90
-max-complexity = 18
+max-complexity = 18
-select = B,C,E,F,W,T4,B9

.flake8_nb

@@ -1,6 +1,5 @@
 [flake8_nb]
-ignore = E203, E266, E402, E501, W503, F403, F401, F841
+ignore = F401, E402, E501
 exclude = .git,__pycache__,.ipynb_checkpoints
 max-line-length = 90
-max-complexity = 18
+max-complexity = 18
-select = B,C,E,F,W,T4,B9

misinformation/display.py

@@ -17,15 +17,18 @@ class JSONContainer:
         return self._data
 
 
-def explore_analysis(image_paths, identify="faces"):
+def explore_analysis(mydict, identify="faces"):
     # dictionary mapping the type of analysis to be explored
     identify_dict = {
         "faces": faces.facial_expression_analysis,
         "text-on-image": text.detect_text,
     }
+    # create a list containing the image ids for the widget
+    # image_paths = [mydict[key]["filename"] for key in mydict.keys()]
+    image_ids = [key for key in mydict.keys()]
     # Create an image selector widget
     image_select = ipywidgets.Select(
-        options=image_paths, layout=ipywidgets.Layout(width="20%"), rows=20
+        options=image_ids, layout=ipywidgets.Layout(width="20%"), rows=20
     )
 
     # Set up the facial recognition output widget
@@ -44,14 +47,18 @@ def explore_analysis(image_paths, identify="faces"):
         output.clear_output()
 
         # Create the new content
-        image_widget.children = (ipywidgets.Image.from_file(image_select.value),)
+        image_widget.children = (
+            ipywidgets.Image.from_file(mydict[image_select.value]["filename"]),
+        )
 
         # This output widget absorbes print statements that are messing with
         # the widget output and cannot be disabled through the API.
         with faces.NocatchOutput():
-            analysis = identify_dict[identify](image_select.value)
+            mydict[image_select.value] = identify_dict[identify](
+                mydict[image_select.value]
+            )
         with output:
-            display(JSONContainer(analysis))
+            display(JSONContainer(mydict[image_select.value]))
 
     # Register the handler and trigger it immediately
     image_select.observe(switch, names=("value",), type="change")

misinformation/faces.py

@@ -72,16 +72,21 @@ retinaface_model = DownloadResource(
 )
 
 
-def facial_expression_analysis(img_path):
+def facial_expression_analysis(subdict):
-    result = {"filename": img_path}
 
     # Find (multiple) faces in the image and cut them
     retinaface_model.get()
-    faces = RetinaFace.extract_faces(img_path)
+    faces = RetinaFace.extract_faces(subdict["filename"])
 
-    # If no faces are found, we return an empty dictionary
+    # If no faces are found, we return empty keys
     if len(faces) == 0:
-        return result
+        subdict["face"] = None
+        subdict["wears_mask"] = None
+        subdict["age"] = None
+        subdict["gender"] = None
+        subdict["race"] = None
+        subdict["emotion"] = None
+        return subdict
 
     # Sort the faces by sight to prioritize prominent faces
     faces = list(reversed(sorted(faces, key=lambda f: f.shape[0] * f.shape[1])))
@@ -120,9 +125,9 @@ def facial_expression_analysis(img_path):
 
     # We limit ourselves to three faces
     for i, face in enumerate(faces[:3]):
-        result[f"person{ i+1 }"] = analyze_single_face(face)
+        subdict[f"person{ i+1 }"] = analyze_single_face(face)
 
-    return result
+    return subdict
 
 
 def wears_mask(face):

misinformation/text.py

@@ -2,9 +2,10 @@ from google.cloud import vision
 import io
 
 
-def detect_text(path):
+def detect_text(subdict):
     """Detects text in the file."""
 
+    path = subdict["filename"]
     client = vision.ImageAnnotatorClient()
 
     with io.open(path, "rb") as image_file:
@@ -14,13 +15,13 @@ def detect_text(path):
 
     response = client.text_detection(image=image)
     texts = response.text_annotations
-    result = {"text": []}
+    subdict = {"text": []}
     for text in texts:
-        result["text"].append(text.description)
+        subdict["text"].append(text.description)
 
     if response.error.message:
         raise Exception(
             "{}\nFor more info on error messages, check: "
             "https://cloud.google.com/apis/design/errors".format(response.error.message)
         )
-    return result
+    return subdict

misinformation/utils.py

@@ -1,5 +1,6 @@
 import glob
 import os
+from pandas import DataFrame
 import pooch
 
 
@@ -29,6 +30,17 @@ def misinformation_prefetch_models():
         res.get()
 
 
+class AnalysisMethod:
+    """Base class to be inherited by all analysis methods."""
+
+    def __init__(self) -> None:
+        # define keys that will be set by the analysis
+        self.mykeys = ["filename"]
+
+    def analyse_image(self):
+        None
+
+
 def find_files(path=None, pattern="*.png", recursive=True, limit=20):
     """Find image files on the file system
 
@@ -55,3 +67,41 @@ def find_files(path=None, pattern="*.png", recursive=True, limit=20):
         result = result[:limit]
 
     return result
+
+
+def initialize_dict(filelist: list) -> dict:
+    mydict = {}
+    for img_path in filelist:
+        id = img_path.split(".")[0].split("/")[-1]
+        mydict[id] = {"filename": img_path}
+    return mydict
+
+
+def append_data_to_dict(mydict: dict) -> dict:
+    """Append entries from list of dictionaries to keys in global dict."""
+
+    # first initialize empty list for each key that is present
+    outdict = {key: [] for key in list(mydict.values())[0].keys()}
+    # now append the values to each key in a list
+    for subdict in mydict.values():
+        for key in subdict.keys():
+            outdict[key].append(subdict[key])
+    # mydict = {key: [mydict[key] for mydict in dictlist] for key in dictlist[0]}
+    print(outdict)
+    return outdict
+
+
+def dump_df(mydict: dict) -> DataFrame:
+    """Utility to dump the dictionary into a dataframe."""
+    return DataFrame.from_dict(mydict)
+
+
+if __name__ == "__main__":
+    files = find_files(
+        path="/home/inga/projects/misinformation-project/misinformation/data/test_no_text/"
+    )
+    mydict = initialize_dict(files)
+    outdict = {}
+    outdict = append_data_to_dict(mydict)
+    df = dump_df(outdict)
+    print(df.head(10))

notebooks/facial_expressions.ipynb

@@ -41,7 +41,10 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "images = misinformation.find_files(limit=1000)"
+    "images = misinformation.find_files(\n",
+    "    path=\"/home/inga/projects/misinformation-project/misinformation/data/test_no_text\",\n",
+    "    limit=1000,\n",
+    ")"
    ]
   },
   {
@@ -64,7 +67,25 @@
   },
   {
    "cell_type": "markdown",
-   "id": "d8067ad1-ef8a-4e91-bcc6-f8dbef771854",
+   "id": "705e7328",
+   "metadata": {},
+   "source": [
+    "We need to initialize the main dictionary that contains all information for the images and is updated through each subsequent analysis:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b37c0c91",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "mydict = misinformation.utils.initialize_dict(images)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a9372561",
    "metadata": {},
    "source": [
     "Next, we display the face recognition results provided by the DeepFace library. Click on the tabs to see the results in the right sidebar:"
@@ -77,7 +98,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "misinformation.explore_analysis(images, identify=\"faces\")"
+    "misinformation.explore_analysis(mydict, identify=\"faces\")"
    ]
   },
   {

notebooks/get-text-from-image.ipynb

@@ -64,6 +64,16 @@
     "    display(Image(filename=i))"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8b32409f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "mydict = misinformation.utils.initialize_dict(mysubfiles)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "07b7a7a3",
@@ -106,7 +116,7 @@
     "    \"\"\"Preprocess the image to enhance features for extraction.\"\"\"\n",
     "    image = cv2.imread(filename)\n",
     "    # preserve the original image\n",
-    "    original = image.copy()\n",
+    "    #     original = image.copy()\n",
     "    # Grayscale, Gaussian blur, Otsu's threshold\n",
     "    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)\n",
     "    # sharpen contrast by first smoothing and then substracting the smoothed and thresholded version\n",
@@ -363,8 +373,7 @@
     "os.environ[\n",
     "    \"GOOGLE_APPLICATION_CREDENTIALS\"\n",
     "] = \"/home/inga/projects/misinformation-project/misinformation-notes/seismic-bonfire-329406-412821a70264.json\"\n",
-    "images = mysubfiles[1:5]\n",
+    "misinformation.explore_analysis(mydict, identify=\"text-on-image\")"
-    "misinformation.explore_analysis(images, identify=\"text-on-image\")"
    ]
   },
   {

requirements-dev.txt

@@ -0,0 +1,11 @@
+deepface
+ipywidgets==8.0.0rc1
+pooch
+retina-face
+opencv-python
+matplotlib
+numpy
+keras-ocr
+tensorflow
+google-cloud-vision
+pytesseract

requirements.txt

@@ -2,10 +2,4 @@ deepface
 ipywidgets==8.0.0rc1
 pooch
 retina-face
-opencv-python
+google-cloud-vision
-matplotlib
-numpy
-keras-ocr
-tensorflow
-google-cloud-vision
-pytesseract