CSV versus API

[susanodd]

There needs to be a Wiki for the CSV routines.

There is documentation on the actual pages, but this is in English. There seems to be some confusion about how to use them and what they do.

PRO CSV:

Three steps

1. Error detection in the input with user feedback.
2. List or "CHANGES" shown to the user prior to carrying out the changes. This allows to user to make sure everything is as wished.
3. Do the changes.

For CSV Update, only the "Signbank ID" (aka Gloss ID), the "Dataset", and the "Field to Update" are required. The other columns should be removed.

There are several interpretations / semantics for "empty" cells. This is highlighted on the CSV page.

For example, an "empty" cell can mean "remove" or "clear" the field. But an empty cell can also mean "do nothing". This second interpretation is useful as it is convenient for "adding additional tags". For example, for the found glosses in a Search result, saved as CSV, then "add the tag" "check phonology" or "missing video", etc.

SENSES

Senses can be CREATED in the CSV. But they currently cannot be "updated".

LEMMAS

Lemmas can be created at the same time a new gloss is created.

Lemma translations can be updated using the UPDATE LEMMA CSV.

Lemmas contain the Dataset in them. Lemmas are SHARED by multiple glosses. That is why GLOSS UPDATE cannot modify the Lemma text because this will have side effects on other glosses.

Updating Lemma translations will have a side effect. The videos will be moved!!!!

GLOSSES

Updating a gloss Annotation will have side effects because the video files will be moved!!!