Save output html as image file

.github/workflows/python-package.yml

@@ -19,21 +19,40 @@ jobs:
 
     steps:
     - uses: actions/checkout@v2
+
     - name: Set up Python ${{ matrix.python-version }}
       uses: actions/setup-python@v2
       with:
         python-version: ${{ matrix.python-version }}
+
+    - name:  Install wkhtmltopdf
+      run:   |
+              if [ "$RUNNER_OS" == "Linux" ]; then
+                  # apt install important_linux_software
+                  sudo apt-get install xvfb libfontconfig wkhtmltopdf
+              elif [ "$RUNNER_OS" == "Windows" ]; then
+                  choco install wkhtmltopdf
+              elif [ "$RUNNER_OS" == "macOS" ]; then
+                  brew cask install wkhtmltopdf
+              else
+                  echo "$RUNNER_OS not supported"
+                  exit 1
+              fi
+      shell: bash
+
     - name: Install dependencies
       run: |
         python -m pip install --upgrade pip
         pip install flake8 pytest
         pip install -r requirements.txt
+
     - name: Lint with flake8
       run: |
         # stop the build if there are Python syntax errors or undefined names
         flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
         # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
         flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
+
     - name: Test with pytest
       run: |
         pytest

ipyplot/_html_helpers.py

@@ -3,20 +3,53 @@
 required for displaying images, grid/tab layout and general styling.
 """
 
+import os
 from typing import Sequence
 
+import imgkit
 import numpy as np
 import shortuuid
 from numpy import str_
 
 from ._img_helpers import _img_to_base64
+from ._utils import (
+    _find_and_replace_html_for_imgkit,
+    _to_imgkit_path)
 
 try:
     from IPython.display import display, HTML
 except Exception:  # pragma: no cover
     raise Exception('IPython not detected. Plotting without IPython is not possible')  # NOQA E501
 
 
+def _html_to_image(html, out_img):
+    if os.path.dirname(out_img) != '':
+        os.makedirs(os.path.dirname(out_img), exist_ok=True)
+
+    html = _find_and_replace_html_for_imgkit(html)
+    options = {
+        # 'xvfb': '',
+        'enable-local-file-access': '',
+    }
+    saving = True
+    while saving:
+        print("Saving output as image under: ", out_img)
+
+        try:
+            # path_wkthmltoimage = r'C:/Program Files/wkhtmltopdf/bin/wkhtmltoimage.exe'
+            # config = imgkit.config(wkhtmltoimage=path_wkthmltoimage)
+            imgkit.from_string(
+                html, out_img,
+                # config=config,
+                options=options)
+        except Exception as e:
+            if "You need to install xvfb" in str(e):
+                options['xvfb'] = ''
+                continue
+            raise
+        saving = False
+
+
 def _create_tabs(
         images: Sequence[object],
         labels: Sequence[str or int],
@@ -25,7 +58,7 @@ def _create_tabs(
         img_width: int = 150,
         zoom_scale: float = 2.5,
         force_b64: bool = False,
-        tabs_order: Sequence[str or int] = None):
+        tabs_order: Sequence[str or int] = None) -> str:
     """
     Generates HTML code required to display images in interactive tabs grouped by labels.
     For tabs ordering and filtering check out `tabs_order` param.
@@ -63,6 +96,11 @@ def _create_tabs(
         By default, tabs will be sorted alphabetically based on provided labels.
         This param can be also used as a filtering mechanism - only labels provided in `tabs_order` param will be displayed as tabs.
         Defaults to None.
+
+    Returns
+    -------
+    str
+        HTML code for class tabs viewer control.
     """  # NOQA E501
 
     tab_layout_id = shortuuid.uuid()
@@ -144,7 +182,7 @@ def _create_tabs(
 
 
 def _create_html_viewer(
-        html: str):
+        html: str) -> str:
     """Creates HTML code for HTML previewer.
 
     Parameters
@@ -235,7 +273,7 @@ def _create_img(
         width: int,
         grid_style_uuid: str,
         custom_text: str = None,
-        force_b64: bool = False):
+        force_b64: bool = False) -> str:
     """Helper function to generate HTML code for displaying images along with corresponding texts.
 
     Parameters
@@ -272,11 +310,14 @@ def _create_img(
     use_b64 = True
     # if image is a string (URL) display its URL
     if type(image) is str or type(image) is str_:
-        img_html += '<h4 style="font-size: 9px; padding-left: 10px; padding-right: 10px; width: 95%%; word-wrap: break-word; white-space: normal;">%s</h4>' % (image)  # NOQA E501
+        matches = ['http:', 'https:', 'ftp:', 'www.']
+        if not any(image.lower().startswith(x) for x in matches):
+            image = os.path.relpath(image)
+        img_html += '<h4 style="font-size: 9px; margin-left: 5px; margin-right: 5px; word-wrap: break-word; white-space: normal;">%s</h4>\n' % (image)  # NOQA E501
         if not force_b64:
             use_b64 = False
             img_html += '<img src="%s"/>' % image
-        elif "http" in image:
+        elif any(image.lower().startswith(x) for x in matches):
             print("WARNING: Current implementation doesn't allow to use 'force_b64=True' with images as remote URLs. Ignoring 'force_b64' flag")  # NOQA E501
             use_b64 = False
 
@@ -309,7 +350,7 @@ def _create_imgs_grid(
         max_images: int = 30,
         img_width: int = 150,
         zoom_scale: float = 2.5,
-        force_b64: bool = False):
+        force_b64: bool = False) -> str:
     """
     Creates HTML code for displaying images provided in `images` param in grid-like layout.
     Check optional params for max number of images to plot, labels and custom texts to add to each image, image width and other options.
@@ -355,7 +396,7 @@ def _create_imgs_grid(
     # create code with style definitions
     html, grid_style_uuid = _get_default_style(img_width, zoom_scale)
 
-    html += '<div id="ipyplot-imgs-container-div-%s">' % grid_style_uuid
+    html += '<div class="ipyplot-imgs-container-div-%s">' % grid_style_uuid
     html += ''.join([
         _create_img(
             x, width=img_width, label=y,
@@ -370,7 +411,7 @@ def _create_imgs_grid(
     return html
 
 
-def _get_default_style(img_width: int, zoom_scale: float):
+def _get_default_style(img_width: int, zoom_scale: float) -> str:
     """Creates HTML code with default style definitions required for elements to be properly displayed
 
     Parameters
@@ -389,13 +430,11 @@ def _get_default_style(img_width: int, zoom_scale: float):
     style_uuid = shortuuid.uuid()
     html = """
         <style>
-        #ipyplot-imgs-container-div-%(0)s {
+        div.ipyplot-imgs-container-div-%(0)s {
             width: 100%%;
             height: 100%%;
             margin: 0%%;
-            overflow: auto;
             position: relative;
-            overflow-y: scroll;
         }
 
         div.ipyplot-placeholder-div-%(0)s {

ipyplot/_img_helpers.py

@@ -33,7 +33,9 @@ def _rescale_to_width(
     return rescaled_img
 
 
-def _scale_wh_by_target_width(w: int, h: int, target_width: int):
+def _scale_wh_by_target_width(
+        w: int, h: int,
+        target_width: int) -> (int, int):
     """Helper functions for scaling width and height based on target width.
 
     Parameters
@@ -56,7 +58,7 @@ def _scale_wh_by_target_width(w: int, h: int, target_width: int):
 
 def _img_to_base64(
         image: str or str_ or np.ndarray or PIL.Image,
-        target_width: int = None):
+        target_width: int = None) -> str:
     """Converts image to base64 string.
     Use `target_width` param to rescale the image to specific width - keeps original size by default.
 

ipyplot/_plotting.py

@@ -7,7 +7,7 @@
 from typing import Sequence
 
 from ._html_helpers import (
-    _display_html, _create_tabs, _create_imgs_grid)
+    _display_html, _create_tabs, _create_imgs_grid, _html_to_image)
 from ._utils import _get_class_representations, _seq2arr
 
 
@@ -88,7 +88,8 @@ def plot_images(
         max_images: int = 30,
         img_width: int = 150,
         zoom_scale: float = 2.5,
-        force_b64: bool = False):
+        force_b64: bool = False,
+        output_img_path: str = None):
     """
     Simply displays images provided in `images` param in grid-like layout.
     Check optional params for max number of images to plot, labels and custom texts to add to each image, image width and other options.
@@ -143,6 +144,9 @@ def plot_images(
         zoom_scale=zoom_scale,
         force_b64=force_b64)
 
+    if output_img_path is not None:
+        _html_to_image(html, output_img_path)
+
     _display_html(html)
 
 
@@ -153,7 +157,8 @@ def plot_class_representations(
         zoom_scale: float = 2.5,
         force_b64: bool = False,
         ignore_labels: Sequence[str or int] = None,
-        labels_order: Sequence[str or int] = None):
+        labels_order: Sequence[str or int] = None,
+        output_img_path: str = None):
     """
     Displays single image (first occurence for each class) for each label/class in grid-like layout.
     Check optional params for labels filtering, ignoring and ordering, image width and other options.
@@ -208,4 +213,5 @@ def plot_class_representations(
         max_images=len(images),
         img_width=img_width,
         zoom_scale=zoom_scale,
-        force_b64=force_b64)
+        force_b64=force_b64,
+        output_img_path=output_img_path)

ipyplot/_utils.py

@@ -1,18 +1,37 @@
 """
 Misc utils for IPyPlot package.
 """
-
+import os
+import re
 from typing import Sequence
 
 import numpy as np
 from PIL import Image
 
 
+def _to_imgkit_path(img_path: str) -> str:
+    # matches = ['http:', 'https:', 'ftp:', 'www.']
+    # if not any(x in img_path for x in matches):
+    #     if '.' in img_path:
+    return "file:///" + os.path.abspath(img_path)
+    # return img_path
+
+
+def _find_and_replace_html_for_imgkit(html: str) -> str:
+    # pattern = r"<img src=\"[/\.].*\""
+    pattern = r"<img src=\"(?!http:|https:|data:|www.|ftp:|ftps:).*\""
+    return re.sub(
+        pattern,
+        lambda x: '<img src="%s"' % _to_imgkit_path(
+            x.group().split('="')[1].replace('"', '')),
+        html)
+
+
 def _get_class_representations(
         images: Sequence[object],
         labels: Sequence[str or int],
         ignore_labels: Sequence[str or int] = None,
-        labels_order: Sequence[str or int] = None):
+        labels_order: Sequence[str or int] = None) -> (np.ndarray, np.ndarray):
     """Returns a list of images (and labels) representing first occurance of each label/class type.
     Check optional params for labels ignoring and ordering.
     For labels filtering refer to `labels_order` param.
@@ -77,7 +96,7 @@ def _get_class_representations(
     return out_images, out_labels
 
 
-def _seq2arr(seq: Sequence[str or int or object]):
+def _seq2arr(seq: Sequence[str or int or object]) -> np.ndarray:
     """Convert sequence to numpy.ndarray.
 
     Parameters

requirements.txt

@@ -5,4 +5,5 @@ bump2version
 pytest
 pytest-cov
 shortuuid
-pandas
+pandas
+imgkit

tests/test_plotting.py

@@ -1,12 +1,14 @@
+import os
 import sys
-from typing import Sequence
+import tempfile
 
 import numpy as np
 import pandas as pd
 import pytest
 from IPython.display import HTML
 from PIL import Image
 
+
 sys.path.append(".")
 sys.path.append("../.")
 import ipyplot
@@ -22,12 +24,16 @@
 ]
 
 BASE_LOCAL_URLS = [
-    "docs/example1-tabs.jpg",
-    "docs/example2-images.jpg",
-    "docs/example3-classes.jpg",
+    "tests/data/21HnHt+LMDL._AC_US436_QL65_.jpg",
+    "tests/data/41edw+BCUjL._AC_US436_QL65_.jpg",
+    "tests/data/100205.jpeg",
 ]
 
+TEST_OUTPUT_IMG = "tests/data/out_img.jpg"
+TEST_OUTPUT_IMG_B64 = "tests/data/out_img_b64.jpg"
+
 LOCAL_URLS_AS_PIL = list([Image.open(url) for url in BASE_LOCAL_URLS])
+LOCAL_URLS_AS_NP = list([np.asarray(img) for img in LOCAL_URLS_AS_PIL])
 
 LABELS = [
     None,
@@ -38,6 +44,7 @@
 
 TEST_DATA = [
     # (imgs, labels, custom_texts)
+    (BASE_LOCAL_URLS, LABELS[1], LABELS[0]),
     (LOCAL_URLS_AS_PIL, LABELS[1], LABELS[0]),
     (BASE_NP_IMGS, LABELS[1], LABELS[0]),
     (pd.Series(BASE_NP_IMGS), LABELS[1], LABELS[0]),
@@ -49,6 +56,19 @@
     (LOCAL_URLS_AS_PIL, LABELS[1], LABELS[1]),
     (LOCAL_URLS_AS_PIL, LABELS[2], LABELS[2]),
     (LOCAL_URLS_AS_PIL, LABELS[3], LABELS[3]),
+    ([os.path.abspath(x) for x in BASE_LOCAL_URLS], LABELS[1], LABELS[0]),
+]
+
+TEST_SAVE_OUTPUT_DATA = [
+    # (imgs, output_img_path, b64, out)
+    (BASE_LOCAL_URLS, "out.jpg", False, TEST_OUTPUT_IMG),
+    (BASE_LOCAL_URLS, "out.jpeg", False, TEST_OUTPUT_IMG),
+    (BASE_LOCAL_URLS, "out/out.jpg", False, TEST_OUTPUT_IMG),
+    (BASE_LOCAL_URLS, "out.jpg", True, TEST_OUTPUT_IMG_B64),
+    (LOCAL_URLS_AS_PIL, "out.jpg", True, TEST_OUTPUT_IMG_B64),
+    (LOCAL_URLS_AS_NP, "out.jpg", True, TEST_OUTPUT_IMG_B64),
+    (BASE_INTERNET_URLS, "out.jpg", True, TEST_OUTPUT_IMG_B64),
+    ([os.path.abspath(x) for x in BASE_LOCAL_URLS], "out.jpg", False, TEST_OUTPUT_IMG),  # NOQA E501
 ]
 
 
@@ -109,3 +129,31 @@ def test_plot_class_representations(
         assert("Ignoring 'force_b64' flag" in captured.out)
 
     assert(str(HTML).split("'")[1] in captured.out)
+
+
+@pytest.mark.parametrize(
+    "imgs, output_img_path, b64, exp_output",
+    TEST_SAVE_OUTPUT_DATA)
+def test_saving_output(
+        imgs, output_img_path, b64, exp_output):
+    with tempfile.TemporaryDirectory() as temp_dir:
+        test_out_path = os.path.join(temp_dir, output_img_path)
+        ipyplot.plot_images(
+            imgs,
+            force_b64=b64,
+            output_img_path=test_out_path)
+
+        assert os.path.exists(test_out_path)
+        assert os.stat(test_out_path).st_size > 10000
+        # assert filecmp.cmp(exp_output, test_out_path, shallow=False)
+
+    with tempfile.TemporaryDirectory() as temp_dir:
+        test_out_path = os.path.join(temp_dir, output_img_path)
+        ipyplot.plot_class_representations(
+            imgs,
+            labels=[0, 1, 2],
+            force_b64=b64,
+            output_img_path=test_out_path)
+        assert os.path.exists(test_out_path)
+        assert os.stat(test_out_path).st_size > 10000
+        # assert filecmp.cmp(exp_output, test_out_path, shallow=False)