English and French DCR (MetaTagger Issue)

Hi,
I have an English and French DCR. What I would like to know is how to write a preprocessor in order to differentiate keywords and description coming from both tabs(English and French). How would the preprocessor recognize/differentiate English/French keywords/description. Is there any documentation you could point me to that deals with this particular issue?

Thank You.

Find more posts tagged with

Comments

Migrateduser

Easiest way:
- create both english and french summary and keyword fields
- create a post-processor that looks at the detected language and copies the right language values to the master field.

More Complex:
- create three content processors: langid, english and french
- the langid only has the transconverter to detect the language
- the english and french each have summary and keyword fields and a postprocessor to rename the language fields (e.g. keywords_fr) to the master field (keyword)
- set up the regexes so that by default all documents go to the langid processor and not the other two
- a final loopback processor looks at the detected language and adds a prefix to the original path and uses the client API to run the document through again.

Jalaja

Hi,
The DCR has both English and French. English and French are within a container. Thanks.

Migrateduser

Where there is mixed language content in a single DCR, you will definitely want to use at minimum two different content processors and probably different metadata fields for each language. Le sommaire francais probably won't help the english user much :}.

For each content processor, you will want to use a pre-processor to remove content from the crackedText that is not appropriate for the targetted language. That will enable the French summarizer to run on the French content and the English summarizer run on the english content.

Cheers,
Clark