When using the indexer's search explorer tool for Azure Cognitive Search the PDF sometimes gets garbled.
For example, if the text in the PDF says "We strive to provide full transparency" the says "We s t r i ve t o p r ov i d e f u l l t r a n s p a r e n c y"
It seems to add arbitrary spaces between letters because in the original document the spacing of letters is increased (however there is actually no "whitespace" character).
It appears their text reader judges the distance between letters, and if it determines the distance is higher than average counts the next letter as a word.
Although this isn't viable because many PDF documents have exaggerated spacing between letters (although no actual whitespace character).
Has anyone ever encountered this type of issue before? How did you resolve it?