Improve the handling of uploaded csv files

Why are we doing this?

We insist on uploaded codesets having no header row, and with the clinical codes in the first column. Both of these likely lead to users opening csv files with Excel prior to uploading, but this can lead to truncation of SNOMED ids. It will improve the user's experience if headers are optional, and the clinical codes can be in any column.

How will we know when it's done?

CSV files with or without headers, and with codes in any column, can be successfully imported into OCL.

What are we doing?

One way would be to:

• upload the csv file
• present the first n rows (10 maybe) to the user in a modal
• allow the user to select the first row of data (either by clicking it, or typing the number in a box)
• allow the user to select the column of clinical codes (again either by clicking it, or typing the number in a box)

Defining delivery tasks guidance

Start by tracing the existing CSV upload and import entry points, then review how uploaded rows and clinical-code columns are currently selected. Define the user flow for previewing the first rows and selecting the data row and code column; done means CSVs with optional headers and codes in any column import successfully into OCL.