LaTeX template refinements

[danielfdsilva]

LaTeX template refinements
Alyssa is finalizing some documentation around current status and what is missing. This will be use to spec out the refinements to be implemented.

subtasks

• PR with documentation: Add documentation #230
• Handle escaping for initial empty LaTeX expression in equation editor. Handle escaping for initial empty LaTeX expression in equation editor. #14
• Fix PDF export in local development environment PDF export is not working in Local development environment #242
• Handle all special characters Support unicode-mathematics and handle all special characters #176
• Inline mathematical equations Inline mathematical equations #180
• Failing gracefully (related to Gracefully handle errors in PDF/HTML generation #177)
• Testing (ensuring serialization works accounting for edge cases, and that it works with the schema provided by the server).
• Architectural improvements serialization infrastructure and dev environment improvements (proposal #1) #248, serialization infrastructure and dev environment improvements (proposal #2) #250, LaTeX pipeline repeatedly downloads figures with wget, outputs pdfs 3x #251

[danielfdsilva]

Relevant issues:

• PR with documentation: Add documentation #230
• Handle escaping for initial empty LaTeX expression in equation editor. Handle escaping for initial empty LaTeX expression in equation editor. #14
• Handle all special characters Support unicode-mathematics and handle all special characters #176
• Inline mathematical equations Inline mathematical equations #180

Other improvements:

• Failing gracefully (related to Gracefully handle errors in PDF/HTML generation #177)
• Testing (ensuring serialization works accounting for edge cases, and that it works with the schema provided by the server).

---

Although not directly related to the latex generator, the whole process could be improved. Currently:

• client gets a serialized json from server
• client puts json in s3 bucket
• serializer converts json to latex and then to html/pdf

We can simplify by having the server put the json in the bucket or sending it directly to the serializer.

[sharkinsspatial]

I agree with @danielfdsilva here #221 that this approach should be simplified. There are a few options each with their own issues and caveats

1. A Lambda which takes the apt id as a parameter, makes the db request for the document and sends it to the serialization ECS task for processing. This is nice since you don't have a service running idle all the time but probably won't work since most of the serialization requests would exceed the API Gateway's 30s request timeout.
2. As @danielfdsilva described running as service we can modify the serialization task to also handle the db request and http request/response and run it as an ECS service on the same instance as the PostgREST container. There may be some difficulty with service port usage and you will need to increase the request timeout configuration for the ELB but this should work as needed.

[guidorice]

Please be sure to see my notes about xelatex and unicode support in #176.