Block Logic

The block normally gets its input from the Microsoft-developed Twitter block, or other blocks with similar output structure. The Twitter block supports the following operations: getLatestPosts, getFriends and getLatestUserPosts. CPB can handle the output of all three operations.

These operations return an array of TwitterStatus instances. Each item holds information of a single post.

For each post, CPB operates the following logic:

1. Examine the post's text found in the text and description members
2. If text member is not defined try title instead, and set it to text
3. Check if any non-English characters exist in the text
   1. If so - filter this post out
4. Invoke the OpenCalais Web service on the post text and parse the results
5. Optionally append a list of identified entities and events to the post text member
6. Add two members, calaisText and calaisTree to the output. These members hold Calais-identified entities and events

The output of CPB is a TwitterStatusWithMetadata object, which is identical in structure to the TwitterStatus object, except for the additional two fields above.

For example, if the input is a TwitterStatus object with an empty description member and the text member set to the following:

NME Radio is now playing Half In Love

The output of the block is a TwitterStatusWithMetadata object with the following members and values:

• text: NME Radio is now playing Half In Love (OpenCalais found: Company - NME Radio [once])

• calaisText: OpenCalais found: Company - NME Radio [once]

• calaisTree: tree with one entity