I suggest to start reading all you can get about text-mining, more specifically possibly mining pubmed articles. We have had related posts here but in the meantime I have learned to understand that, if beginning from scratch, it can be difficult to even know the right terms to search for in google.

I could be a good start to simply go through all questions on Biostar on text mining: http://www.biostars.org/search/?q=text-mining (72 hits). Then you can possibly better define your goals and set priorities. You need to have a certain amount of basic knowledge about the topic to be able to define your project correctly, and it is certainly helping to look at what others are working with.

Likely a good start:

• http://www.biostars.org/post/show/48265/text-mining-clinicaltrialsgov-for-drug-gene-interactions/#48280
• http://www.biostars.org/post/show/12326/examples-of-full-text-mining-on-the-pubmed-central-open-access-subset/

For searching pubmed, you'll want to read and understand about eUtils. I'm not clear on what you are trying to do other than to find information about drugs, but pubmed searching sounds like it will be part of the workflow.