-
- Making this setup of automated generation of Power Point presentations is possible
- using the ***python-pptx*** library, for which parameters need to be defined,
- coordinates to place our objects of interest and their sizes need to be calculated.
- For more information on this library, see: https://pypi.org/project/python-pptx/
-
+
+
Automating the generation of Power Point presentations with our layout is done
+ using the python-pptx library.
+ For more information on this library, see:
+ python-pptx (Pypi)
''', unsafe_allow_html=True)
-
+ st.divider()
st.markdown('''
- ## Parameters defined to insert the images on the slides
+ ##
Parameters to define
- The specific coordinates of all the elements desired in the slides were previously tested
- on a pptx by manually arranging the images with the intended number of rows, columns,
- images per slide, labels, text boxes and a size adequate to gain enough insight of the
- results while keeping the number of slides as low as possible. Once all the intended
- content for a single slide was set into approximate position, all the sizes were
- calculated and set to precise numbers in order to make it reproducible and iterable.
+
- ''')
+ The specific coordinates (in centimeters) for all the desired elements in the
+ slides were previously tested by manually arranging images with the intended number
+ of rows, columns, images per slide, labels, and text boxes. The size of the elements
+ was adjusted to provide sufficient insight into the results while minimizing the
+ number of slides. Once the approximate position of all the content on a single slide
+ was set, precise measurements were calculated and applied to ensure reproducibility
+ and ease of iteration.
+
+
+
+ ''', unsafe_allow_html=True)
# Diplay the images of the slide coordinates
- col_1_row_1, col_2_row_1 = st.columns([1, 1], gap="small")
- with col_1_row_1:
+ a, img1_container, b = st.columns([0.5, 9, 0.5], gap="small")
+ with img1_container:
st.image(image="https://github.com/EdRey05/Streamlit_projects/raw/main/002_Automated_PPTX_PLA/Automated_PPTX_goal.jpg",
- caption="Desired slide layout", use_column_width=True,)
- with col_2_row_1:
+ caption="Desired slide layout (Title = experimental condition, Subtitle = Image, Image name = ROI name, P = particle count)",
+ use_column_width=True, )
+
+ st.markdown('''
+
+
+ Once the images and text of interest were arranged, the coordinates for each space
+ that the objects would fit into were measured as follows. NOTE: That required
+ to resize the images and few look stretched in an axis due to their original aspect
+ ratio (this can be customized for other experiments depending on the shape and aspect
+ ratio of the images).
+
+
+
+ ''', unsafe_allow_html=True)
+
+ c, img2_container, d = st.columns([0.5, 9, 0.5], gap="small")
+ with img2_container:
st.image(image="https://github.com/EdRey05/Streamlit_projects/raw/main/002_Automated_PPTX_PLA/Automated_PPTX_coordinates.jpg",
- caption="Coordinates for each desred object", use_column_width=True)
+ caption="Coordinates for each desired object (x, y)",
+ use_column_width=True,)
st.markdown('''
+
+
All the parameters illustrated above are the following:
- 1. Slides of 16:9 ratio (34cm width, 19cm height, all measurement units set to cm).
-
- 2. Two titles at the very top, their text boxes of 17cm width and 1.5cm height each,
- side by side with no separation from the top corners. Title bold font, subtitle normal
- font, and both Times New Roman size 32 points).
-
- 3. Below the title+subtitle, there is a 0.6cm space to the first row of images, so the
- first row starts at X, 2.1cm (X = distance from the left edge).
-
- 4. Each image will be resized to 3.25cm width by 3cm height. The pairs consist of a
- fluorescence image on the left side and a particle analysis on the right side, with no
- space separating them.
-
- 5. There are 0.25cm separating each pair (5 pairs per row, 4 pairs per column, total
- 20 pairs/cells = 40 images per slide), and also 0.25cm separating the first and last
- pair of a row from the edges of the slide.
-
- 6. There is a text box right under each image (the borders touch). For the fluorescence
- image the text refers to the name of the ROI analyzed, whereas the particle image shows
- the count of particles for that particular cell so we can quickly inspect whether all
- the coloured particles coincide with the actual puncta we see in the fluorescence image
- and the final count (or if there was an error due to the quantification process). This
- text boxes are 3.25cm width by 1cm height, contain Times New Roman text size 20 points.
-
- 7. The second row is separated 0.3cm from the bottom edge of the labels mentioned in 6.
- Same for the third and fourth row (which almost ends at the very edge of the slide).
-
- **Note:** The image and label coordinates are declared in a list of tuples (coordinates
- can't change for this notebook), so we can iterate through the images grouped per slide
- (1-20) and only use the required coordinates (we don't need to iterate 20 times all the
- time if the slides only have few images, it is better to iterate per image in the slide
- and use the index to find its corresponding coordinates). The filling order will be left
- to right, top-down. Since the fluorescence and particle images both come from the same
- cell, their names will be "1_2.jpg" and "1_1.jpg", respectively, so the tuple of
- coordinates consists of 4 elements: the first is the distance from the left edge of the
- slide for the fluorescence image, the second is the distance from the top edge of the
- slide for the fluorescence image, the third is the distance from the left edge of the
- slide for the particle image, and the fourth is the distance from the top edge of the
- slide for the particle image. This way, we can use the same ROI name to get both images
- by replacing part of the directory and the "_2.jpg" for "_1.jpg", however, the index in
- the for loop will be the same, since we have the coordinates of both images in the same
- element of the list (tuple of 4 coordinates).
-
- ''')
+
+
+
+ 1. Slide ratio: Use a 16:9 ratio (34cm width, 19cm height - all measurements
+ are given in cm).
+
+ 2. Title and subtitle: Two titles at the top, side by side inside a text box
+ of 17cm width, and 1.5cm height (right at the top corners). The title has bold Times
+ New Roman font, size 32 points, whereas the subtitle has normal Times New Roman font,
+ size 32.
+
+ 3. Titles separator: Below the title and subtitle text boxes, there is a 0.6cm
+ vertical space to the first row of images.
+
+ 4. Image size: Each image is resized to 3.25cm width by 3cm height. The pairs
+ (yellow + green rectangles) come from the same cell (region of interest). The
+ fluorescence image is on the left side and the particle mask image is on the right
+ side, with no space separating them horizontally.
+
+ 5. Image labels: There is a 3.25cm width by 1cm height text box right under
+ each image. The text is normal Times New Roman font, size 20 points. The text of the
+ fluorescence image indicates the name of the .roi file quantified for that image (so
+ we can find it for any purpose). The text of the particle mask image indicates the
+ count of particles for that particular region of interest. We should evaluate whether
+ the number matches what we see in the fluorescence image, and whether there are
+ more/less coloured and black (uncounted) particles in the mask than expected.
+
+ 6. Image pair separation: There are 5 pairs of images per row, 4 pairs per
+ column = 20 pairs of original+quantified images (cells or regions of interest). All
+ pairs of images are separated by other pairs both horizontally and vertically by 0.25cm.
+
+ 7. Image filling order: The images are filled left to right, top-down. They
+ appear in the csv file sorted alphabetically, but we inserted them into the slides in
+ natural sorting order (more intuitive).
+
+ 8. Additional details: Due to the size of the slide and the indicated sizes and
+ coordinates, the last row finishes very close from the bottom of the slide, and the
+ last column finishes very close from the right side of the slide (may not be visible as
+ the text boxes have no fill and the right side picture has white background).
+
+
+
+
+ **Notes:**
+
+ * Since the fluorescence and particle mask images come from the same cell/region of
+ interest, their names are the same but have a different number aft the underscore:
+ "100_2.jpg" and "100_1.jpg", respectively.
+
+ * The specific image and label coordinates are hardcoded in tuples (see snippet below).
+
+ * The tuple of coordinates consists of 4 elements: the first is the distance from the
+ left edge of the slide for the fluorescence image, the second is the distance from the
+ top edge of the slide for the fluorescence image, the third is the distance from the
+ left edge of the slide for the particle image, and the fourth is the distance from the
+ top edge of the slide for the particle image. This way, we can use the same ROI name to
+ get both images by replacing part of the directory and the "_2.jpg" for "_1.jpg",
+ however, the index in the for loop will be the same, since we have the coordinates of
+ both images in the same element of the list (tuple of 4 coordinates).
+
+
+
+ ''', unsafe_allow_html=True)
+
+ snippet = '''
+ # All coordinates are stated always in the same order: From left first, from top second.
+
+ # Title text box dimensions and coordinates (centimeters)
+ title_width = 17
+ title_height = 1.5
+ title_left_coordinate = 0
+ title_top_coordinate = 0
+
+ # Subtitle text box dimensions and coordinates (centimeters)
+ subtitle_width = 17
+ subtitle_height = 1.5
+ subtitle_left_coordinate = 17
+ subtitle_top_coordinate = 0
+
+ # Size and coordinates for the 20 pairs of images (centimeters)
+ image_width = 3.25
+ image_height = 3
+ image_coordinates = [
+ (0.25, 2.1, 3.5, 2.1), (7, 2.1, 10.25, 2.1), (13.75 , 2.1, 17, 2.1), (20.5, 2.1, 23.75, 2.1), (27.25, 2.1, 30.5, 2.1),
+ (0.25, 6.4, 3.5, 6.4), (7, 6.4, 10.25, 6.4), (13.75, 6.4, 17, 6.4), (20.5, 6.4, 23.75, 6.4), (27.25, 6.4, 30.5, 6.4),
+ (0.25, 10.7, 3.5, 10.7), (7, 10.7, 10.25, 10.7), (13.75, 10.7, 17, 10.7), (20.5, 10.7, 23.75, 10.7), (27.25, 10.7, 30.5, 10.7),
+ (0.25, 15, 3.5, 15), (7, 15, 10.25, 15), (13.75, 15, 17, 15), (20.5, 15, 23.75, 15), (27.25, 15, 30.5, 15)
+ ]
+
+ # Size and coordinates for the 20 pairs of text labels (centimeters) (+3cm top coordinate of images)
+ image_labels_width = 3.25
+ image_labels_height = 1
+ image_labels_coordinates = [
+ (0.25, 5.1, 3.5, 5.1), (7, 5.1, 10.25, 5.1), (13.75 , 5.1, 17, 5.1), (20.5, 5.1, 23.75, 5.1), (27.25, 5.1, 30.5, 5.1),
+ (0.25, 9.4, 3.5, 9.4), (7, 9.4, 10.25, 9.4), (13.75, 9.4, 17, 9.4), (20.5, 9.4, 23.75, 9.4), (27.25, 9.4, 30.5, 9.4),
+ (0.25, 13.7, 3.5, 13.7), (7, 13.7, 10.25, 13.7), (13.75, 13.7, 17, 13.7), (20.5, 13.7, 23.75, 13.7), (27.25, 13.7, 30.5, 13.7),
+ (0.25, 18, 3.5, 18), (7, 18, 10.25, 18), (13.75, 18, 17, 18), (20.5, 18, 23.75, 18), (27.25, 18, 30.5, 18)
+ ]
+ '''
+ e, snippet_container, f = st.columns([1, 8, 1], gap="small")
+ with snippet_container:
+ st.code(snippet, language="python")
st.write("---")
st.markdown('''
- ## Potato
+ ##
Overview of the processing strategy
+
+
- * Unzip the "Data.zip" file to make a "Data" folder, over which we can iterate/walk
+ 1. Unzip the "Data.zip" file to make a "Data" folder, over which we can iterate/walk
through.
- * Once the folder is ready, look for the csv results file of each experimental
- condition (1 per folder/condition), in which we have a collection of most of the
- information we need.
+ 2. Once the "Data" folder is ready, get all the folders inside as these are the
+ experimental conditions or groups. Each
+
+ 3. For each Group folder, there must be a Group/Quantification/results.csv file.
- * We know that the number of csv's found is the same number of experimental conditions
+ 4. We know that the number of csv files found is the same number of experimental conditions
we need to iterate through. We get their paths, names, and we already know the folder
- structure where the images are located: **content->Data->ExpCondition->Cropped cells->** ,
+ structure where the images are located: **Data/ExpCondition/Cropped cells/** ,
which contains 3 subfolders: **Fluorescence, FM_Particles, and T_Particles**
- * Once we know the number of experimental condition folders, we will iterate through any
- number of them to extract the information of all the images analyzed.
+ 5. Once we know the number of experimental condition folders, we will iterate through any
+ number of them to extract the information of all the images.
- * Since the csv files already contain almost all the information we need (subtitle, ROI
- names, T particle could and FM particle count), we will iterate through the rows of the
+ 6. Since the csv files already contain almost all the information we need (subtitle, ROI
+ names, T particle count and FM particle count), we will iterate through the rows of the
csv file instead of walking through the directory of fluorescence images (we could also
extract the info from the path but we would need multiple steps to split different
sections of the path). This strategy also allows us to easily convert the ROI names into
- integers so we can sort them properly (natural sorting, the ouput of the quantification
- script is not in the correct order).
+ integers so we can sort them properly (natural sorting, the csv file is not in this
+ order).
- Now we have a huge list containing all the info for all the images uploaded by the user.
- The next step is to group these images, following a few rules:
+
+
- 1. The main idea is to pass 20 images to the function that makes the slides.
+ Up to this step, we will have a huge list containing all the info for all the images
+ uploaded by the user. The next steps are to group these images to know which ones go
+ together into the same slide, following a few rules:
+
+
+
- 2. We start with the first experimental condition (title of slide), first subfolder of
+ * The main idea is to pass 20 images to the function that makes the slides.
+
+ * We start with the first experimental condition (title of slide), first subfolder of
the original image (like Row_01_05, which is the subtitle), and then we see how many
cells/ROIs were quantified there.
- 3. If we have exactly 20 (unlikely), we just make one slide. If we have less than 20, we
+ * If we have exactly 20 (unlikely), we just make one slide. If we have less than 20, we
also make one slide and leave empty spots. If we have more than 20 images, we take the
first 20, make one slide, take the next (up-to) 20 using the same title and subtitle,
make a new slide and so on, until we don't have more images in that subfolder.
- 4. To accomplish this strategy, we need to iterate through the list generated above
+ * To accomplish this strategy, we need to iterate through the list generated above
(all_info_for_slides), in which all the information for all the images of all the
conditions is side by side. Because of that, we can iterate through the same level of
the list, take the info of the current item/element, and add its content to a new
temporary variable (list) with 20 spots available. Before adding the info of the current
item/element, we check whether the 20 spots have been filled, check whether the title
has changed, and check whether the subtitle has changed. Any of those 3 cases triggers
- the jump to a new slide so we have to pass the grouped image info from the temporary
- variable to a final variable, clear the temporary variable content and then add the
- current image info to the new slide group. Finally, we repeat this process over and over
- until we have checked all the items/elements, and the final variable, product of this
- loop, will contain all the image information separated/grouped by images that will go
- together in a single power point slide.
-
- 5. With the final variable we will be able to iterate through each element=slide, call
- the function that makes the slides, and pass the current element information to insert
- the images for that slide.
-
- ''')
+ the immediate jump to a new slide so we have to pass the info of the grouped images
+ from the temporary variable to a final variable, clear the temporary variable content
+ and then add the current image info to the new slide group. Finally, we repeat this
+ process over and over until we have checked all the items/elements, and the final
+ variable, product of this loop, will contain all the image information grouped by
+ slide.
+
+ * With the final variable we will be able to iterate through each element (slide), call
+ the function that makes the slides, and pass the current information of the images to
+ insert to the slide. We do this over and over until we have made a slide for all groups
+ of images specified, and finally we save the presentation. We do this twice, once for
+ the images with Thresholding, and once for the images with Find Maxima (info for both
+ sets of images and counts is contained in the same variable all_info_for _slides.
+
+
+
+ ''', unsafe_allow_html=True)
return
###################################################################################################