I’ve been using Tap Forms for a few years as a digital log book of patients. The primary purpose is to capture billing details. Prompted by watching Sam’s excellent video on
OCR Receipt Extraction. I’d like to have a script to auto populate fields with the OCR from a scan of the patient’s ID label.
.
My current workflow is to scan a patient’s hospital label with my iPhone, transfer it to the photo field in a new TF record before manually entering data from the scanned image. It’s the digital equivalent of sticking a label in a book in order to annotate it with other data like start and finish times, items, etc.
.
What I’d like to do is scan the label with my iPhone (I use Scanner Pro which does a pretty reliable OCR), create a new record in TF that contains the image in a photo field and the OCR’d text in a notes field, then use a form script to automatically populate fields in the TF form. Because the data is sensitive I want to keep it encrypted, only on my own devices and synced using near field only.
.
Let me show an example.
.
The same fictional patient on different hospital labels gives OCR text:
.
HEELER, MR Bandit
MRN: OT000123456
DOB: 04/09/1978 Age: 44 (M)
10 CATTLEDOG ROAD
BETOOTA HEIGHTS 4483
Home Ph: 0491 577 426
MC: 2123456791
S/Net No:
DVA:
BETOOTA PRIVATE
Adm No
CT04589023
Adm Date: 12/01/24
13:48
Att Dr: Fixem, Ayecan S
Ref Dr: Payne, Ophelia M
Fund: HFHHF
50006000
MC Exp: 05/2028 MC Ref: 1
Pen No:
0038708Y
AD: 12/01/2024
Adm No. 4589023
HEELER, MR Bandit
10 Cattledog Rd
Betoota Heights, QLD 4483
PH: 0491577426 M: 0491577426
MC: 2123456791/1 Exp:05/2028
CL: PI HHF/T 50006000
PC NIL
AD: 12/01/2024
FIXEM, DR AYECAN
UR: 123456
Adm No. 4589023
DOB: 04/09/1978 (44Y)
MAR: M
Gender: Male
BET 0038708Y
.
The bits of data I need are first name and surname, DOB, MC, MC Ref, fund code and fund number (‘HHF’ and 50006000 in the example).
.
These would map to TF fields surname (text), first name (text), DOB (DD/MM/yyyy), MC (number), MC ref (number), fund code (link from form) and fund number (text; sometimes contains letters) in a single TF form called ‘Cases’.
.
The stickers are rectangular and in landscape orientation, so they have information in two columns. The lines of OCR text aren’t always in the same order, e.g. sometimes the admission number or MRN appear between the address lines. The identifier snippets are always present and predictable, though. For instance, fund code is always found after ‘Fund:’ or ‘CL: PI’ which will always be on the same line.
.
Sam’s script for his fuel receipts is awesome. I want to do something like it, but I also want to know how it works so I can fix it and build on it. I’m new to scripting and I’m trying to get simple things to work before adding more complexity. Not having much success though. Here’s my attempt:
.
function mc() {
var mc_id = ‘fld-abbabb14632f41718ffb4a259cebe63a’;
var label_text_id = ‘fld-d80875c6c53048c8bcf97b11e190118d’;
var input = record.getFieldValue(label_text_id);
var mcRegex = /[2-6]{1}[0-9]{9}/;
var found = input.match(mcRegex);
if (record.getFieldValue(mc_id)) {
return;
}
record.setFieldValue(mc_id, found);
console.log(found);
document.saveAllChanges
}
mc();
.
It finds mc in the notes field text but doesn’t populate the MC field.
.
Can anyone help me with this?
.
I have a couple of related questions after watching Sam’s video:
.
1. How can I create a new record in TF on iOS that contains both the scanned image (jpg) and OCR text string of the scan in a notes field? I’m able to get one piece of information in easily using Shortcuts/x-callback, but two?
2. How can I create a link from form using scripting? I’d like to extract the fund code (‘HHF’ in the example; Fund:\s?HF?(\D{3})|CL: PI\s?(\D{3})) then link from the HHF record in the ‘Funds’ form.
.
