Some text unexpectedly turned into images by the PDF Extract API

[Garth22367887bhae]

Hi,

I've been using the PDF Extarct API to display the content of a PDF as an HTML file, so far it's working well, but I'm seeing a couple of odd cases that you might be interested in. The following happens in a pdf that has sets of fill-in-the-blank questions.

For the folllwing text from the PDF:

SELECT c.name 'Channel',
COUNT(i.dataid) 'Count of Items'
FROM __________ i, __________ c
WHERE i.parentid = c.___________
AND i.____________ = 208
GROUP BY c.name, i._____________ ORDER BY COUNT(i.___________) DESC

The last line is extracted as an image instead of text, see below

[Garth22367887bhae]

Here is the json that gets generated:

    {
      "Bounds": [
        231.8408966064453,
        331.61090087890625,
        435.8748016357422,
        364.3813934326172
      ],
      "Font": {
        "alt_family_name": "Courier New",
        "embedded": true,
        "encoding": "Identity-H",
        "family_name": "Courier New",
        "font_type": "CIDFontType2",
        "italic": false,
        "monospaced": false,
        "name": "EDGRTV+CourierNew,Bold",
        "subset": true,
        "weight": 700
      },
      "HasClip": false,
      "Lang": "en",
      "Page": 24,
      "Path": "//Document/L[17]/LI/LBody/P[3]",
      "Text": "WHERE i.parentid = c.___________ AND i.____________ = 208 ",
      "TextSize": 10.5,
      "attributes": {
        "LineHeight": 12.5
      }
    },
    {
      "Bounds": [
        231.8199005126953,
        319.11590576171875,
        435.860107421875,
        339.3914031982422
      ],
      "Font": {
        "alt_family_name": "Courier New",
        "embedded": true,
        "encoding": "Identity-H",
        "family_name": "Courier New",
        "font_type": "CIDFontType2",
        "italic": false,
        "monospaced": false,
        "name": "EDGRTV+CourierNew,Bold",
        "subset": true,
        "weight": 700
      },
      "HasClip": false,
      "Lang": "en",
      "Page": 24,
      "Path": "//Document/L[17]/LI/LBody/P[4]",
      "Text": "GROUP BY c.name, i._____________ ",
      "TextSize": 10.5,
      "attributes": {
        "LineHeight": 12.625,
        "SpaceAfter": 6.625
      }
    },
    {
      "Bounds": [
        231.8199005126953,
        306.6208953857422,
        448.57875061035156,
        326.8963928222656
      ],
      "Page": 24,
      "Path": "//Document/L[17]/LI/LBody/Figure",
      "attributes": {
        "BBox": [
          231.8309999999983,
          311.41999999999825,
          445.9019999999873,
          322.2189999999973
        ],
        "Placement": "Block"
      },
      "filePaths": [
        "figures/fileoutpart45.png"
      ]
    },

[Garth22367887bhae]

And here is a screenshot of what it looks like in the PDF:

[Joel Geraci]

Can you share the actual PDF. I'd like to use it to train our AI.

[Garth22367887bhae]

Sorry, It's a client document that I can't share.