Handle documents the import could not read

After a large import, some documents produce no events: a scanned letter, a photo of a signed page, a fax. A few may not have been processed at all. Before you tell the team the chronology is complete, every one of them needs a decision, so nothing is silently missing.

Two questions decide what to do with each document. Is it a scan or a photo? Then it needs OCR, so its text can be read. Does it hold dated events at all? A price list or a CV may not, and can stay as it is.

Files that could not be processed

1. When the import finishes, the Unprocessable files list opens if any file could not be read. It gives each file's Title and, under Description, the reason. No events were created from these files.

2. Write down the files before you close the list, because it does not open again. Click OK on a row once you have noted it.

3. Get each file again in another format, for example as a PDF, and upload it to the timeline again.

Documents without events

1. Click the red badge in the toolbar on the left: Events couldn't be found on some documents. Click here to decide how to handle them. The number on it is the number of documents.

   

2. The documents are listed. Click Show to open one and see what it is. Tick the documents you want to treat the same way, or tick Select all.

3. For scans and photos, click OCR and retry. Choose the language the documents are written in under Select Language: OCR reads the text in that language. Then choose:

   • Auto Proceed puts the events found straight on the timeline. Leave it off to review each document's events before they are added.
   • Split PDFs is for a file that holds several documents, such as a scanned bundle.

   Click retry. The documents leave the list while they are scanned and searched for events again.

   

4. For documents where you would rather mark the events yourself, such as a handwritten note or a plan with dates on it, click manual extraction. The first document opens for review. Select text and click Add event, or click Mark a date and draw a box around a date on a scanned page. Accept or remove every event on the document, then click Next document to go on, and Finish review after the last one.

If you stop halfway, the documents you have not finished wait behind another badge in the toolbar: Documents await your attention. Click here to review and take necessary action.

Tips from practice

• Check after every import. Look at the toolbar on the left before you report the chronology complete. A red badge there is a document that has not been dealt with yet.
• One language per retry. OCR reads in the language you choose. In a bundle with letters in two languages, retry the documents for each language separately.
• Poor text in a PDF. Some PDFs have a text layer that reads badly, from an old scan. When you upload them with JuristIQ as the processing mode, turn on Force OCR Scan under Provide extra instructions, so they are scanned again rather than read as they are.
• Bundles. If a single scan holds a whole bundle, it is easier to work with once it is split into its documents. See split a scanned bundle into separate documents.

Back to the import itself: build the chronology from the case documents. For a scan you want to search rather than turn into events, see make scanned documents searchable.