MD parsing libs are unable to identify the tables in the markdown version of the word

[praveenrv]

Hi Everyone,
The markdown version of the word for the tables ( blocks or regular tables ) isn’t in the expected format by markdown parsing libraries ( one of them is commonmark-java)
Can I get suggestions on how to parse this markdown for tables?? My use case is to generate a structured JSON of the word document with all the assets references, so I started parsing the MD format of the word.

For example:

MD gave by Franklin for metadata block:
+----------------------------+
| Metadata |
+-------------+--------------+
| Title | MD Test |
+-------------+--------------+
| description | MD Test page |
+-------------+--------------+

Expected format by the library:
| Metadata | | |-------------|--------------| | Title | MD Test | | description | MD Test page |

[chrischrischris]

To summarize from internal slack thread:
Franklin uses gridtables: https://github.com/adobe/remark-gridtables

The only guaranteed stable format is the html, so that should be what's parsed. This is documented here:
adobe/helix-html2md#6