I'm struggling to figure out if a TM/TU match is a context match, and the documentation isn't helping.

The request object posted to the GroupShare API does not offer any option to tell GroupShare something about the current
context. However, the response from that same call can deliver several different attributes that contain the word 'context', but at the moment I can't see how to make use of these to identify whether our current match is a context match.

The pertinent information we would guess is this:

"results": [
        {
            "memoryTranslationUnit": { ... }
            "translationProposal": { ... }
            "cascadeEntryIndex": -1,
            "contextData": {
                "leftSource": 2440472997293638732,
                "leftTarget": 3798523
            },
        ...

In the 'leftSource' and 'leftTarget' attributes there are hash values that we would guess are calculated hashes of the preceding segments. If this is true then we could simply calculate the hash of the preceding segment ourselves, compare it with the hash value and than we would know if it is a context match. Which raises the questions:

• Is this assumption correct?
  • If no: how can we determine if a match is a context match?
  • If yes: how can we calculate the hash? The documentation says that the hash value is not even always the hash of the preceding segment, e.g. in cases where there is no preceding segment.
  • If yes: there should be the possibility that a match has several different contexts (several different preceding segments). How would that be reflected in the response?