Overview

Try to learn rust by writing a basic spreadsheet engine: rustsheet

Features

• single sheet (no multi-sheet workbooks)
• data stored column-wise in vectors of cells
  • number of columns/rows grows and shrinks as data are added and removed
  • columns grow or shrink by adding new or taking away column vectors
  • rows grow or shrink by increasing or decreasing the length of all column vectors
  • all column vectors are maintained at the same length and grow/shrink together
  • cells each hold a single value with one of a few possible data types
    • Int (i32)
    • Real (f64)
    • Text (String) → a fun addition would be two separate string cell types where one is just a string and the other is a formula string that can operate on other cells
    • null (no value) no need to explicitly store null cells, just store cells with actual values
• sheet state is stored in a json file managed by this program (sheet.json)
• command-line interface for modifying state
  • call signature something like rustsheet <subcommand> args ...
  • <subcommand> could be one of a few options
    • read_cell <loc> - print the value of a cell at <loc> to stdout
    • write_cell <loc> <value> - write <value> into a cell at <loc>, prints updated sheet to stdout
    • read_sheet - print the entire contents of the sheet to stdout, first line is "<n_cols>, <n_rows>" then each line after that is " " for all non-null cells ordered by column, row
    • count_rows - print the number of rows in the sheet to stdout
    • count_cols - print the number of columns in the sheet to stdout
    • add_row - add a new row to the sheet (just increases the n_rows attribute of Sheet), prints new number of rows to stdout
    • add_col - a new column to the sheet, prints new number of columns to stdout
    • shrink - reduce the number of rows and columns in the sheet to the minimum amount necessary to hold all non-null cells, prints new dimensions as "<n_cols> <n_rows>" to stdout
    • delete_cell <loc> - delete the cell (if any) at loc, prints the updated sheet to stdout
  • subcommands that return information (e.g. read_cell, count_rows) just print the result to stdout, others (e.g. write_cell, add_row) just modify the sheet state (in sheet.json)
  • indexing (<loc>) is in typical <upper_letter><number> format where the letter portion denotes the column (A, B, C, ...) and the number denotes the row (1, 2, 3, ...)
• BONUS: write up a simple Python/tkinter GUI app for viewing/interacting with the spreadsheet engine (via CLI on the backend)

sheet state

sheet state stored in sheet.json with the following layout`

Instead of using json just use a simpler text format (in sheet.txt) for now with the following layout

<n_cols> <n_rows>
<cell_loc> <cell_val>
<cell_loc> <cell_val>
... ...

This is the same formatting as what is printed to stdout when the read_sheet subcommand is used.