NGWi · October 27, 2024 16:52
diff --git a/Tanach-projectToDos b/Tanach-projectToDos
 **Frontend:**
 Update screenshot to show text at top, verse text, and ... Roots section (:-D!!!)
 In verse page, words should not get pushed down to fit under Text, they should jump to the next line. Also they should stay aligned right even during resizing, and extra lines should be aligned right as well.
 Add back button to VersesShow that makes API call to words/:id. Pull first verse that contains the word besides the current verse in VersesShow when the button under the word is clicked.
 The verses in words' show should be a span of the words in the verse so the matched_word can be highlighted and hyperlinked.
 Basic Tailwind styling, header, footer.
 The verses in words' show and verses' show should be hyperlinked. In the words' show only the book:chapter:verse should be hyperlinked.
 Verses page should have layout of x01.htm, or switch all pages to format of /p/pt/pt01.htm (see Translation Strategy).
 Home page can have search boxes for book:chapter:verse, or word, and a button for file ingestion.
 Wheel and spoke diagrams of related words, possible construction in back-end of entire word table as such.

 **Backend:**
 There should be a "if doesn't exist" for verse and word just like matches even though slows it down so the database stays ordereded
 Add check in parsing algos for pre-existing verses and words too. (Also check that Match check works.)
 Have to filter out `--` without replacing with " "  like `-`. This could be the reason the occasional word has a space afterwards, which needs .trim() on the frontend to keep it flush right.
 Think of efficent way of serving fuzzy matches (one off, e.g., letter added, removed, or swapped.) Is another table necessary?
 Create actions should have logic to ingest uploaded file and move seeds.rb logic to "fat models."
 Make it easy to to optionally, or by default, suspend Match flow until after upload all docs.
 Create node-tree of the data model for posting on read-me.
 Optimize search algo.
 Build more robust tests.

 **Translation Strategy:**
 Switch all pages to format of /p/pt/pt01.htm
 Ingest translation to translation_verses with common index of verse_id (even though book:chapter:verse will curtail overflow of rare alternate verse breaks, will have to work them out either way)
 Use ML to identify most likely word to translation_word matchups and test.
 Use ML to find k-means cluster of translations of forms of that word starting with longer words.
 Find 3-letter combinations that were never words and try to deduce the meaning.
 Go all the way down to one-letter words.

 **Full-Stack Features:**
 Enable Route of verses/book:chapter:verse.
 Enable route of words/word.
 Release on Render.io.
 Possible usage through users using the engine to ingest texts with limited storage size and time.
 Progress bar that updates, e.g., every chapter. Should display approx. progress of count=Words.all.count; sqrt(word.id/count)
	Frontend:
	Update screenshot to show text at top, verse text, and ... Roots section (:-D!!!)
	In verse page, words should not get pushed down to fit under Text, they should jump to the next line. Also they should stay aligned right even during resizing, and extra lines should be aligned right as well.
	Add back button to VersesShow that makes API call to words/:id. Pull first verse that contains the word besides the current verse in VersesShow when the button under the word is clicked.
	The verses in words' show should be a span of the words in the verse so the matched_word can be highlighted and hyperlinked.
	Basic Tailwind styling, header, footer.
	The verses in words' show and verses' show should be hyperlinked. In the words' show only the book:chapter:verse should be hyperlinked.
	Verses page should have layout of x01.htm, or switch all pages to format of /p/pt/pt01.htm (see Translation Strategy).
	Home page can have search boxes for book:chapter:verse, or word, and a button for file ingestion.
	Wheel and spoke diagrams of related words, possible construction in back-end of entire word table as such.

	Backend:
	There should be a "if doesn't exist" for verse and word just like matches even though slows it down so the database stays ordereded
	Add check in parsing algos for pre-existing verses and words too. (Also check that Match check works.)
	Have to filter out `--` without replacing with " " like `-`. This could be the reason the occasional word has a space afterwards, which needs .trim() on the frontend to keep it flush right.
	Think of efficent way of serving fuzzy matches (one off, e.g., letter added, removed, or swapped.) Is another table necessary?
	Create actions should have logic to ingest uploaded file and move seeds.rb logic to "fat models."
	Make it easy to to optionally, or by default, suspend Match flow until after upload all docs.
	Create node-tree of the data model for posting on read-me.
	Optimize search algo.
	Build more robust tests.

	Translation Strategy:
	Switch all pages to format of /p/pt/pt01.htm
	Ingest translation to translation_verses with common index of verse_id (even though book:chapter:verse will curtail overflow of rare alternate verse breaks, will have to work them out either way)
	Use ML to identify most likely word to translation_word matchups and test.
	Use ML to find k-means cluster of translations of forms of that word starting with longer words.
	Find 3-letter combinations that were never words and try to deduce the meaning.
	Go all the way down to one-letter words.

	Full-Stack Features:
	Enable Route of verses/book:chapter:verse.
	Enable route of words/word.
	Release on Render.io.
	Possible usage through users using the engine to ingest texts with limited storage size and time.
	Progress bar that updates, e.g., every chapter. Should display approx. progress of count=Words.all.count; sqrt(word.id/count)