More Data From the PDF#488
Conversation
# Conflicts: # server/api/lib/mavat/index.js # server/api/model/plan.js # server/package-lock.json # server/package.json # server/tests/tables_structs/plan_struct.js
| // this function look for the correct columns for the given headers, and returns a factory embedded with these findings | ||
| const rowAbstractFactoryChart1_7 = (firstPageOfTable, headersStartIndex) => { | ||
| // clean the headers | ||
| firstPageOfTable = firstPageOfTable.map(row => row.map(cell => cell.replace(/\n/g, ' ') |
There was a problem hiding this comment.
what happens if firstPageOfTable is undefined or not an array?
There was a problem hiding this comment.
It's always an array. We can add a check beforehand, but we firstPageOfTable comes from the PDF extension and it's consistent.
Adding a check here will require adding the same check on at least 10 places.
| contains: getFromArr(row, containsIdx), | ||
| scale: getFromArr(row, scaleIdx), | ||
| number_of_pages: getFromArr(row, numberOfPagesIdx), | ||
| edit_date: getFromArr(row, editDateIdx), |
There was a problem hiding this comment.
yup, for the db. By using snake casing we don't need to change the map's keys when inserting into the db.
|
|
||
| if (headersStartIndex === -1) { | ||
| // it's an error | ||
| log.error('didn\'t find headers for some char 1.6'); |
There was a problem hiding this comment.
can you please add planId to the log so we can assign to a process for investigation?
There was a problem hiding this comment.
Passing a planId here requires me to pass planId into more software layers (at least 2 more) and change at least 25 lines in at least 15 different files. It will be a bit complicated.
The crawler that runs in production doesn't log the planId if it fails. Can you check the logs to see if we got a lot of "didn't find headers for *" to see if it worth investing in?
|
|
||
| if (headersStartIndex === -1) { | ||
| // it's an error | ||
| log.error('didn\'t find headers for some chart 3.1_without_change'); |
There was a problem hiding this comment.
Same thing as the comment on the log.error on 1.6
2 participants
This PR brings us tables:
This PR also brings an improvement in runtime since it knows to cut the appendixes from a heavy PDF.