Open Scriptures

In a recent post to the Open Scriptures mailing list, it was suggested that we use md5 (or another cryptographic hash) to generate unique IDs for each token (a “token” is the fundamental unit of text (most often a word) in our API database models). Today we discussed the implementation of this on IRC, and it was fairly stimulating.

First of all, md5 is broken and deprecated, due to possible collisions (two different pieces of data can result in the same hash). Since we will be dealing with millions of tokens, we decided not to test our luck, unlikely though a problem may be. SHA-256 has no known collisions, so we decided it was best to use that algorithm.

SHA-256 is implemented in Python’s standard library hashlib, so that is good. For exapmle:

>>> import hashlib
>>> hashlib.sha256("Hello world!").digest()
'\xc0S^K\xe2\xb7\x9f\xfd\x93)\x13\x05Ck\xf8\x891NJ?\xae\xc0^\xcf\xfc\xbb}\xf3\x1a\xd9\xe5\x1a'

Needless to say, such a digest would not be very good for use in a RESTful URI scheme. So, hashlib also offers a hexadecimal option:

>>> hashlib.sha256("Hello world!").hexdigest()
'c0535e4be2b79ffd93291305436bf889314e4a3faec05ecffcbb7df31ad9e51a'

That is still not the best, since that makes for a very long string. So, we have the option of using base64 encoding:

>>> import base64
>>> base64.b64encode(hashlib.sha256("Hello world!").digest())
'wFNeS+K3n/2TKRMFQ2v4iTFOSj+uwF7P/Lt98xrZ5Ro='

That is shorter, but it includes the “/” character, which is a no-no for URI design. Luckily base64 includes a function for this exact purpose:

>>> base64.urlsafe_b64encode(hashlib.sha256("Hello world!").digest())
'wFNeS-K3n_2TKRMFQ2v4iTFOSj-uwF7P_Lt98xrZ5Ro='

So that is safe for URIs, but being case-sensitive and having ambiguous characters makes it not the best for working with. So, base32 to the rescue:

>>> base64.b32encode(hashlib.sha256("Hello world!").digest())
'YBJV4S7CW6P73EZJCMCUG27YREYU4SR7V3AF5T74XN67GGWZ4UNA===='

There you have it: shorter than hex, and easier to work with than base64. If we end up using cryptographic hashes an token unique IDs, I’m pretty sure this is how we’ll do it.

The discussion around using cryptographic hashes as unique identifiers in our models is ongoing. Essentially we need to decide how best to make a unique hash for each token. Please join us in #openscriptures on irc.freenode.net with any input.

At the moment the Open Scriptures project is working on developing an internet API for querying scriptural text and metadata. The basic task is to create “a common API for many datasets.” However, before the API can be implemented, the underlying relational database models must be established. To that end, Weston has been working on implementing database models for the API using Django (the development platform for Open Scriptures).

One of the most challenging aspects has been finding out how to record structural information about the text: verses, chapters, title headings, etc. There’s also been a desire to not rely on any particular structural marker in the database’s organization. So the base unit for storing the text is not a chapter or verse, but what is called a “token”. A token is comprised by one of the three atomic structures of a text – word; punctuation; whitespace. Of course, there may be cases where even the basic token can be split, but you’ve got to start somewhere.

To provide structure, Weston has written a token linkage system, where you can define a certain structure (e.g. “Verse 12″) and, using the features of a relational database, connect it to the tokens which should be included in that structure. There is even a feature for non-linear token linkages, if anyone finds a use for that.

Another piece of the puzzle is deciding how to express various types of metadata about the text in the database. One important type of metadata is the parsing information for Greek or Hebrew tokens. That parsing information could be provided in a simple string (e.g. “verb PAI3S”) which client applications would then have to interpret based on established conventions. This is not ideal, however, since it would seriously hamper the querying power of the API. It is best to instead use a database model for parsings. The challenge here comes in supporting multiple languages. Once again, relational database features will assist with this problem, since we can assign one Greek or Hebrew (or any other language) parsing to each token’s metadata. If there is a difference of opinion on a parsing, we can even store multiple parsings for each token.

I am optimistic about the potential of this project. Once the API is nailed down, there will be a lot of great opportunities for “client” apps, using whatever framework they wish. Until then, the API has to be finalized and garnished with built-in methods, and the models have to be tested with real data (which requires that the data be ported to the models in the first place). This is where we can use help from all sorts of people, from Python programmers to database experts to linguists and biblical scholars. It’s a good time to be interested in the scriptures and open source software.