Educational Archive for "wristsaver v5.5"

Please note, that this project is obsolete, and is no longer maintained. No additional documentation will be released. No additional installation support will be provided.

Please note, that user font creation process, demonstrated in this readme file, is an extremely unpractical and tedious endeavor. More information regarding this project is available at the following address: GitHub pages.

What is This

Item	Description
What is this?	A project, saved into git repository, hosted at GitHub website.
What does it contain?	Several python tools for printing handwritten documents. Minimal demo pseudofont, purely for demonstration purposes, not suitable for actual document printing.
Whom are they for?	For beginner python developers.
What is the current project status?	✔️Functional. 😴Archived, outdated, no longer maintained.
Why is it obsolete?	The code is unnecessarily obscure and rigid — introducing any meaningful changes would require complete project restructuring. User font creation is complex and, taking into account the possibilities of modern machine learning tools, requires a prohibitive amount of manual labor.
What is the purpose of this repository?	Host an educational archive with working code.
Why is it named like that?	Because at one point in time it allowed to save several human wrists from unnecessary injuries.

Installation

For the best experience, please use the specific version of the Spyder python IDE. It can be installed via pip install spyder==5.5.6 command.

1. Clone this repository in a directory of choice.
2. Navigate to this directory.
3. Follow this manual in order to:
   1. Create the virtual environment
   2. Activate this virtual environment
   3. Make the Spyder work in this virtual environment.
4. Install the dependencies, listed in the requirements.txt, into this virtual environment.
5. In Spyder, open the menu option "Projects"->"New Project...", select "Existing directory", set "Location" to the dir of the cloned repository and click "Create".
6. Open the menu option "Tools"->"Preferences", select "Run", set "Working directory settings" to "The current working directory" option.
7. Make sure that current working dir is the cloned repository dir.

How to print a handwritten document

1. Open the step5_Printer.py file with Spyder.
2. Scroll to the bottom.
3. Adjust the my_font_folder variable.
4. Choose the menu option "Run"->"Run".

The .png images, named consequtively 1.png, 2.png, 3.png... (their number depends on the amount of text being printed and user font size) will start appearing in the cloned repository folder. The user may want to manually post-process them via any available image editor.

The example output with a fine-tuned font could be as follows (please note that for privacy reasons the original output image, which was 2480x3500px and 300 dpi resolution, had to be cropped and severely scaled down):

Customize the text that will be printed

1. Open the giant_lecture.html file.
2. Carefully edit its contents (make sure all tags are closed, etc).
3. Keep the tag structure as simple as possible — only headers, paragraphs and new lines are recognized. _{Probably. I don't remember.}

Please note, that theoretically wristsaver can print in any language — any letter, number, text symbol, emoji, pictogram, word, phrase, anything. In order to achieve this in practice, the user needs to manually add each desired custom item into his font and provide appropriate metadata. The process will be described in further sections.

Customize output image size, resolution, etc

1. Open the main.css file.
2. Carefully edit its contents (make sure all tags are closed, etc).
3. Keep the values as simple as possible.

How to create a user font

For demonstration purposes, a minimal demo font is included at all development stages — from raw to printable. The contents of this font are sampled from the freely available Kaggle handwriting dataset, available at the official website.

Please see the expected folder structure (some files are omitted):

■ cloned repository directory
├─■ fonts (root folder for all fonts)
│ ├─■ step0_raw_kaggle_test_images.zip
│ ├─■ step1_manually_split_kaggle_test_letters.zip
│ ├─■ step2kaggle_rem_bg_create_df.zip
│ ├─■ step3kaggle_write_image_width_and_height_to_df.zip
│ └─■ step4kaggle_Editor (folder with actual printable font)
├─■ step2_rem_bg_create_df.py
├─■ step3_write_image_width_and_height_to_df.py
├─■ step4_Editor.py
├─■ step5_Printer.py (already explained in the previous section)
└─■ step10_Marker.py

[Step 0] Obtain raw image with handwriting

1. Scan a page with some handwritten text on it.
2. Create a folder fonts/my_font_0raw.
3. Put a disposable copy of the scanned image into the folder.

[Step 1] Split this image into desired fragments

Please note, that the desired fragments can have any shape or size — no restriction. Any differences will be mitigated via metadata editing tool in Step 4.

Let's say that during Step 0 the user has obtained raw.png image. Let's assume that this raw.png contains the following handwritten text: The weather is good 100% :D. Let's assume that the user wants to extract the following sub-images from this big raw.png:

• The image, containing word The
• The images for each individual letter (T,h,e, ,w,e,a,t,h,e,r, ,i,s, ,g,o,o,d, ,1,0,0,%, ,:,D)
• The image, containing the frequenty-used number 100
• The image, containing the smile :D.

The means to achieve this are entirely up to the user. Worst-case scenario is manual image editing: select desired region, save it to disk.

Let's assume that at this moment the user has a lot of small images that contain the desired extracts from the big raw.png.

1. Create a folder fonts/my_font_1split.
2. Create a folder fonts/my_font_1split/letters_lowercase.
3. Create a folder fonts/my_font_1split/letters_uppercase.
4. Create a folder fonts/my_font_1split/numbers.
5. Create a folder fonts/my_font_1split/my_custom_category.
6. Create a folder fonts/my_font_1split/it_can_be_named_anything.
7. Put disposable copies of some of the small images into the appropriate subfolders.

Please note, that the small images can be placed anywhere — in root fonts/my_font_1split, in some user-defined subfolder. The target image location is entirely up to the user.

Step 2 — Remove Backgrounds and Create Metadata

1. Duplicate the folder fonts/my_font_1split.
2. Rename it to be fonts/my_font_2metadata.
3. Open the step2_rem_bg_create_df.py file with Spyder.
4. Scroll to the bottom.
5. Adjust the folder_with_letter_images variable to the fonts/my_font_2metadata.
6. Choose the menu option "Run"->"Run".

The images in the fonts/my_font_2metadata folder will be edited inplace — their background will become transparent. Additionally, df.xls files will be created.

Step 3 — Automatically Get Some Metadata

1. Duplicate the folder fonts/my_font_2metadata.
2. Rename it to be fonts/my_font_3metadata.
3. Open the step3_write_image_width_and_height_to_df.py file with Spyder.
4. Scroll to the bottom.
5. Adjust the folder_with_letter_images variable to the fonts/my_font_3metadata.
6. Choose the menu option "Run"->"Run".

The df.xls files will be edited inplace.

Step 3.3 — Add Text Labels to Metadata

1. Duplicate the folder fonts/my_font_3metadata.
2. Rename it to be fonts/my_font_printable.
3. Open one df.xls in this folder with any supported editor.
4. Add new column named text.
5. Input the desired text (example: t_uppercase.png has text T, a_version_slanted.png has text a, 1.png has text 1, smiley_face.png has text 😊, word_the.png has text ^).

Please note that only one-character symbols are supported. _{Probably. I don't remember.}

At this moment in time the font is officially usable: it has enouh metadata. The printed results will most likely be odd — letter spacing is incorrect, tall images are placed wrong, etc.

Step 4 — Adjust Metadata via GUI Editor

1. Open the step4_Editor.py file with Spyder.
2. Scroll to the bottom.
3. Adjust the my_font_folder variable.
4. Choose the menu option "Run"->"Run".
5. Two GUI windows will show up:

6. Click on any letter in the GUI window with letters.
7. The letter will become highlighted by red border, and the lines will appear:

8. Use W,A,S,D to adjust the letter's position relative to its neighbouring letters.
9. When satisfied, press ENTER
10. In the Font Browser GUI window press the Save changes button.
11. In the Font Browser GUI window press the Re-render button.
12. Choose another letter to edit.

Press the ` key to toggle different editor modes: the "default" one and the "relative to the previous letter" one.

After this, attempt to print the document via step5_Printer.py again, and see if the results are better.

The user may further enhance the font by using the step10_Marker.py editor.
The user may further enhance the font by manually adding the previous letter values into the df.xls files.

License

The actual licenses are available at the beginning of each source code file; superficial overview:

• The GUI elements that rely on the PyQt5 library are generally subject to GPL v3.
• Modules that don't rely on the PyQt5 library are generally subject to Apache 2.0.

Any unmarked file is implied to be subject to Apache 2.0.