Recover Context Match from a competitor's TMX file

I have a vendor returning this TMX from Memsource:

<tu>
<tuv xml:lang="en">
<prop type="context_prev">Browse Hardware Manuals</prop>
<prop type="context_next">Benchtop Measurement and Test</prop>
<seg>PC-Based Systems</seg>
</tuv>
<tuv xml:lang="zh-cn" creationdate="20210503T164558Z" changedate="20210503T164558Z">
<prop type="project">blah</prop>
<prop type="domain">R&amp;D</prop>
<prop type="subdomain">Product R&amp;D</prop>
<prop type="created_at">1620060358000</prop>
<prop type="created_by">xxxx</prop>
<prop type="modified_at">1620060358000</prop>
<prop type="modified_by">xxxx</prop>
<prop type="filename">Raw text to be translated 05032021.xlsx</prop>
<seg>基于PC的系统</seg>
</tuv>
</tu>

I want to turn this into a TMX that GroupShare will understand fully, including the context.  I know GS only cares about "previous" in CM, and its TU look like this:

<tu creationdate="20210503T164558Z" creationid="xxxx" changedate="20210503T164558Z" changeid="xxxx" lastusagedate="20210503T164558Z">
<prop type="x-LastUsedBy">xxxx</prop>
<prop type="x-Context">xxxxxxxxxxxxxxxxxx, 28116399300057872</prop>
<prop type="x-ContextContent">Browse Hardware Manuals | | 浏览硬件手册 | </prop>
<prop type="x-Origin">TM</prop>
<prop type="x-ConfirmationLevel">ApprovedSignOff</prop>
<tuv xml:lang="en-US">
<seg>PC-Based Systems</seg>
</tuv>
<tuv xml:lang="zh-CN">
<seg>基于PC的系统</seg>
</tuv>
</tu>

My question is this.  It looks like GS uses some sort of HASH in the <prop type="x-Context"> to tell the system what the previous segment is.  But, since relatively recently, it also uses a clear text at <prop type="x-ContextContent"> with source and target separated by pipes.

  • Is SDL willing to give us their HASH algorithm, so I can convert the Memsource TMX to the Trados version?
  • Do I even need the hash? Would creating a TMX with <prop type="x-ContextContent"> sufficient?
  • Is the target string really important in the <prop type="x-ContextContent"> to define context?
  • Is there any other method to recover the context?  Is there a difference if I import the TMX through Studio vs importing it directly in GroupShare?

Parents Reply Children
  • Thanks for letting us know we don't need the Hash.

    this is probably irrelevant anyway since the memsource TMX does not provide previous source AND target and that's what's required.  Not just source.

    I think there is a way to solve this. I can easily find the translation for that previous source by looking it up in the same Memsource TMX. All I have to do is look for that string as segment source and make sure that "context_next" is the original source text.

    So I think I have what I need to do the conversion from Memsource to GroupShare and keep the context match.

    By the way, we discovered a strange behavior.  When you import a TMX into GroupShare using Trados Studio 2017 (to work around a bug with the word boundary finder DLLs, support Case Details - 00417738), it creates context match information (we can tell by re-exporting the TMX from GroupShare), and our best guess is that it uses the order of the TUs in the TMX as basis for context. Was that a bug that was then fixed in later version of Studio?