couchdb-python-utils/doc_utils.py

This script facilitates validating JSON and Python datastructures, and then POST or PUTing them to a CouchDB server.

This is tested for the 0.8x branch, but not 0.9x. It seems there are some small breaking changes (or perhaps just bugs ;) in 0.9, so the script may require some minor modifications before using it there.

Usage

This script facilitates validating JSON and Python datastructures, and then POST or PUTing them to a CouchDB server.

The script has one required and one or more optional positional parameters:

python doc_utils database-name
python doc_utils database-name path/to/folder/or/file
python doc_utils database-name doc1.py doc2.py doc3.json doc4.js

If you specify a folder, then the script will examine all .py, .json, and .js files in the folder, and--if they meet the format requirements--send them to CouchDB.

Invoke help at the command-line for more options

python doc_utils --help

Format for Individual Documents

For format for a .py file a dictionary.

{
    '_id': '123123',
    'title': "hi"
}

The format for .js and .json files is a JS dictionary.

{
    "_id":"101",
    "desc":"and a description"
}

In general the distinction between the two will typically be quite small, but remember that JSON technically doesn't permit a comma after the last element in an array/dictionary.

• Individual documents are always PUT.
• Thus they must have an ID specified.
• They will not overwrite existing documents with the same ID.

Format for Bulk Documents

The format for bulk documents is identical to that of individual documents, but they must be a list/array instead of a dict.

Python looks like:

[
    {'_id': '123123','title': "hi"},
    {'title': "goodbye"}
]

Actually, thats exactly what the JSON version will look like as well. Cool, eh?

• Bulk documents are always POSTed, and thus do not require IDs (but may have them).
• This will overwrite existing documents with the same IDs.