You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add to documentation description about what SAS Url is – link to service docs about training (we might already do this) Changes in docs [Form Recognizer] #12216
For Copy, hard to find resource id and region. Add docstrings to clarify this more. Changes in docs [Form Recognizer] #12216 Add in the documentation how/where to get the region for Copy and the list of available regions
Add documentation to clarify getFieldValue() and getValueText() (note: not python naming) - Deferred discussion to Architecture board. Consider if renaming fields to labelData or labelTextInformation should help ?Tracked in[formrecognizer] more renames to allow for non-text elements #12274
The text was updated successfully, but these errors were encountered:
Add documentation to clarify getFieldValue() and getValueText()
Order fields in position for the document (unsupervised map of fields). Currently, field ordering is not maintained.
Helper function to search and return fields to shortcut getFields().get("FieldName") (Might need more discussion).
Helper/accessor to get Table cell data more easily (Table indexer or getRow() like Js has)
[Service follow up] Cannot understand the breaking for FormLines. For a line "* = Exceeds Maximum: Indicates the service exceeds patient/group benefit year maximum." why is it breaking on "=" or "*"
Models exposed to users to have toString() method for formatting purposes.
Add readme code snippet pointers for statements in key concepts.
For Copy, hard to find resource id and region. Add docstrings to clarify this more.
Consider adding a mix of fromUrl and stream overload code snippets in the main readme.
Consider adding a sample showing an example for converting a recognized form to known predefined fields. [receipt sample]
Consider moving TrainingDoucments errors to top level training result (needs more discussion)
DEBRIEF
Data returned must not be sorted - model data to match the order of the form
clarify what receipt is on documentation
begin_recognize_receipt rename to **_form_stream vs **_from_uri - or use a helper method to read from uri
value vs value_text in receipt result was not intuitve (consider renaming?)
recheck copy api documentation
LAYOUT API is ambigious - should the onfo be included in the recognized form results?
highlight what fields are available in the receipt returned
valueText vs fieldText vs labelText - naming confusion
Add in the documentation how/where to get the region for Copy and the list of available regions
text
vstext_content
– clarify ref docs Changes in docs [Form Recognizer] #12216To follow up:
getRow()
or similar not natural to have agetCells()
and diving into the cell information to find the row it belongs to. Also, it could use a better data structure here. [constant pain point to get to the field] - Follow up (Mariana/Anne) Table indexer issue - [Design] [Form Recognizer | Table Indexer] Implement indexer on ExtractedTable to index into Cells array azure-sdk-for-net#9975 and Tracked in [formrecognizer] consider allowing indexing into table for recognize_content #10894fromUrl
and that should have worked. Circling back to if we should consider adding an overload that takes in a file input. - (Waiting to know Johan's thoughts, Issue - [FEATURE REQ] Consider adding overload to take a Local file input for recognize methods azure-sdk-for-java#12402 (comment)) Tracked in [formrecognizer] support container local file paths #12276Deferred discussion to Architecture board. Consider if renaming fields to
labelData
orlabelTextInformation
should help ? Tracked in [formrecognizer] more renames to allow for non-text elements #12274The text was updated successfully, but these errors were encountered: