Provide commands to create database file [Feature Request]

[chrisgulvik]

Can you provide example commands to create a database?

I'd like to make a super small db for built-in functionality testing of the software within a container. For example, just adding 2 different species genomes (perhaps just 1 Gram Pos here and 1 Gram Neg here) and querying with 1 of them. This would also speed up any dev work you do and make GitHub Actions a possibility for you. Nothing fancy needed for commands to share, just notes in shell would be helpful.

If you added an example on how to create a GTDB database, that would also take the workload off of you, ensure future utility of the pkg as taxonomy continues to change, and avoid future issues like #40 and #43 .

[chrisgulvik]

I just saw @jfy133 also ran into this same need here. It looks like simply making a dmd db isn't enough, so any tips?

[jfy133]

It might just be that particular genome being used during testing is not in the database (I'm not 100%) sure, so at least instructions would indeed be helpful :)

[fullama]

I could make a small db for testting yeah.. the thing is that the database versions are tied to a gunc version so opening it up for users to make their own database might cause confusion. maybe it would be fine, i dont know tbh. I could add a small test db in the next version release.. would an addition gunc test command work for ye maybe?

[jfy133]

Thanks for looking into this @fullama !

For clarity, the purpose for me (and I guess @chrisgulvik ) is not to allow users to make a custom database - a small test DB would indeed suffice.

A small database that is compatible with the standard gunc commands rather than a specific gunc test would be perfect, as being as close to simulating a 'real' GUNC run, the better. E.g., ultimately I want to make sure that the wrapper of the standard GUNC commands used in 'real world' applications functions correctly - I just don't need the full database to do so.

So a separate gunc test command would not help here (for me) - as the test I want to make is of the normal GUNC commands themselves.

[fullama]

OK no prob.. ive added it to the milestones for the next release.. ill update here when its ready

[krishn-cpu]

Install SQLite in the container (if not already available)
apt-get update && apt-get install -y sqlite3

Create a SQLite database and a table for genome entries

sqlite3 test_genomes.db <<EOF
-- Define a table with species name, Gram stain type, and genome data
CREATE TABLE genomes (
id INTEGER PRIMARY KEY,
species_name TEXT NOT NULL,
gram_stain TEXT NOT NULL,
genome_data BLOB NOT NULL
);
EOF

Insert a Gram-positive bacterial genome (Bacillus subtilis as an example)
sqlite3 test_genomes.db <<EOF
INSERT INTO genomes (species_name, gram_stain, genome_data)
VALUES ('Bacillus subtilis', 'Positive', 'GENOME_DATA_HERE');
EOF

Insert a Gram-negative bacterial genome (Escherichia coli as an example)
sqlite3 test_genomes.db <<EOF
INSERT INTO genomes (species_name, gram_stain, genome_data)
VALUES ('Escherichia coli', 'Negative', 'GENOME_DATA_HERE');
EOF

Query the database to verify inserted records
sqlite3 test_genomes.db <<EOF
-- Print all genome records
SELECT * FROM genomes;
EOF

[jfy133]

@krishn-cpu can you clarify? Is this an example work around? Could you maybe provide a more 'real world' example (I'm not sure what you mean by GENOME_DATA_HERE for example?