Where should the content of web scrap be stored? In the previous version, it is stored in desc. If we keep doing so, we cannot simply tell the type of snippets based on its fields and an extra field such as snippet_type needs to be added.

Refactor into declarative object-storage mapping models

Since in git storage, one user's data is stored independently, we might also store users' data on ES independently for each user. ES supports searching across different indices, so regarding searching and sorting it doesn't matter much. The only issue is that latest ES requires index and type to be the same. With ES version such as version 5.6.5 (deployed for testing platform), it still supports different index and type.

The current plan is to only keep tag name in es, discarding desc, private, init_time fields that are present for tags in the user space. This approach will lose the capability to do some stats on user data, such as new tags this year.

Such as tag list, create a tag, get a tag, update a tag, delete a tag. Tag deletion and update will also update snippets that have this tag.

Build basic restful apis for Snippet and Tag with django-rest framework. The goal is to accomplish the full chain of restful api development with django-rest, yaml files, git and es. So probably only typical data api for Snippet and Tag will be implemented at this stage.

Advanced ones, such as merge tags

Previously tags are all stored in one yaml file as multiple documents. However, this causes difficulty in telling change of tag from diff. Since one yaml file per tag is still quite human readable, a decision is made to switch the design to using one yaml file per tag.

Thinking of using a set of new terms for Knowledge Engine: "Mind" is filled with "Snippets". Snippets are rich snippets, with urls and multimedia attachments. Snippets can be classified as various types depending on its fields, such as bookmarks, web scraps... The classification is not mandatory and is a bit arbitrary. It is up to the apps built on top of snippets to decide whether to use it or not. Some snippets are more experiences in nature, and significant experiences can be used to be built up knowledge that is core to one's understanding of a domain.

To uses, for example, snippet id.

In addition to CRUD, also add and remove children and change ordering.

Add the initial test cases for snippet and tag

If the ordering of these documents can be ensured, then diff line number might be used to determine which document gets updated (for simple cases such as only one document changed. If multiple document change, it would be more challenging). For deletion and addition, it is relatively easy.

Get snippets by tag, various searches, bulk edit, comments and annotations.