Change the NCBI Entrez search to reduce test failures

[peterjc]

Often the continuous integration tests are failing here:

esearch -db nucleotide -query "its1 AND Phytophthora[Organism] AND 150:800[Sequence Length]" | efetch -format fasta

Either NCBI Entrez is timing out and failing directly, or apparently it is returning with partial results, and the DB count checks are not always satisfied (currently we say we expect at least 100 taxonomy table entries).

Probably simplest to add a limit to the Entrez search string, and also a bash if statement checking the number of records saved to the FASTA file?

[peterjc]

Even the reduced search can fail, this as 12:25 GMT/UTC, https://circleci.com/gh/peterjc/thapbi-pict/131

$ esearch -db nucleotide -query 'its1 AND Phytophthora[Organism] AND 150:800[Sequence Length] AND 1900:2000[Publication Date]' | efetch -format fasta
WebEnv value not found in search output - WebEnv1 
Db value not found in fetch input
Exited with code 255

[peterjc]

Avoiding this by checking in the sample file, done as part of #69 to get the CI tests to pass.