Skip to content

Latest commit

 

History

History
99 lines (65 loc) · 2.98 KB

meetingnotes.md

File metadata and controls

99 lines (65 loc) · 2.98 KB
Raw

MPropulator Meeting notes

10/23/2015 - Kickoff

RSF agreed that at first, this would be a tool for programmers. In the future they would like it to implement a GUI for researchers.

They drafted an initial implementation to look something like this:

import sys
sys.path.append('N:/library_folder')
import MPropulator

# Timestamp on file name
# default argument of timestamp = true to indicate whether timestamp would be included in file name
MPropulator.create_config(config="config.csv", tab_names_same_as_csv = True) # optional shell for config file
MPropulator.populate(shell="shell.xlsx", output="out.xlsx", config="config.csv", timestamp=True)

ExcelPopulator.create_config would create a semi-filled in Excel file that the programmer would then complete. See config.csv for an example. At their next meeting, they agreed to divy up the tasks.

10/26/2015

import ExcelPopulator import ExcelPopulator.populate

-- Library MPropulator | init.py | create_config.py | populate.py | tab.py

10/27/2015

Tab can have multiple tables - tabname and table name column Table class and Tab class

Tab: list of tables, method is write_all Table: write_table

Check if tables overlap -- raise an error to the user if so

1 workbook, 1 config

create_config function parse_config populate

Franco:

  • Justify MPR_Workbook
  • Populate

Sam:

  • Parse_config
  • Tab class

Ray:

  • skip_rows
  • Table functions

11/4/2015

(Ray accidentally deleted his original copy of the notes for this meeting; these are from memory.)

Ray, Sam, and Franco agreed that populate() was getting too large and that it should be modularized. They agreed that the function should be split up into roughly these different ones:

  • Validate populate() input
  • Read in config file
  • Validate config file

F showed R and S a little bit of SAS's ability to run python through the command prompt. F said that Mpropulator's next feature should be to support this by creating a function (or modifying populate()) that writes tables without a config file so that SAS programmers can use it more easily. S noted that another feature should be reading directly from SAS datasets and writing to Word documents.

Also, Ray needs to document col_to_number.

11/11/2015

Franco used split instead of str.split for performance reasons and map instead of a for loop so that the code would take advantage of C.

Check if skiprows is list using an assert.

Is there a path extractor function so that we can avoid using franco's sepPath function?

Look into pandas for way to deal with missing values Also do aprseConfig['skipcols'] = map(helpers.col_to_number)

Check that each element in skipcols is a letter Convert cols in skipcols to numbers where we convert skiprows

Ray will write tests for populate, write_tabs, helpers F will write tests for readConfig, init S will test validations

Create bad_configX.csv with different types of errors

Sam can look into concatenating errors into one big thing.