remove requirements-kg2-build.txt entries that are no longer needed now that ontobio is removed

[saramsey]

bmt pymongo

[saramsey]

Try this:

HTMLParser==0.0.2
jsonlines==3.0.0
PyMySQL==0.9.3
PyYAML==5.4
requests==2.31.0
snakemake==5.5.4
xmltodict==0.12.0
validators==0.15.0

[bazarkua]

I verified all packages in RTX-KG2/setup/requirements-kg2-build.txt using: for name in find . -name \*.py | grep -v misc-tools | grep -v archieve | grep -v neo4j; do python3 $name; done > ../run-log2.out 2>&1 and then looking for import errors or warnings.

[ecwood]

I'm concerned that some of the removed packages are actually needed. In particular, bmt is the Biolink Model Toolkit, which is used for validation purposes. graphviz is used by snakemake to create the build process diagram (ex. the Snakemake Diagram here). I believe that jsonpickle is how we download pickle, which is used by a couple of ETL scripts.

[saramsey]

Where is bmt used in the RTX-KG2 project code? I did a search both via GitHub's web browser UI and of the master branch code, like this:

sramsey-laptop:Desktop sramsey$ git clone [email protected]:RTXteam/RTX-KG2.git
Cloning into 'RTX-KG2'...
remote: Enumerating objects: 14409, done.
remote: Counting objects: 100% (1958/1958), done.
remote: Compressing objects: 100% (283/283), done.
remote: Total 14409 (delta 1743), reused 1676 (delta 1675), pack-reused 12451 (from 3)
Receiving objects: 100% (14409/14409), 4.28 MiB | 8.11 MiB/s, done.
Resolving deltas: 100% (9355/9355), done.
sramsey-laptop:Desktop sramsey$ cd RTX-KG2
sramsey-laptop:RTX-KG2 sramsey$ find . -type f -exec grep bmt {} /dev/null \;
sramsey-laptop:RTX-KG2 sramsey$

and did not get a hit.

[saramsey]

Same goes for jsonpickle:

sramsey-laptop:RTX-KG2 sramsey$ find . -type f -exec grep jsonpickle {} /dev/null \;
sramsey-laptop:RTX-KG2 sramsey$

maybe we are pulling them in via some indirect PyPI package dependency?

[saramsey]

And graphviz is installed via apt-get, I believe.

[saramsey]

And pickle is a built-in python package, is it not?

[bazarkua]

And pickle is a built-in python package, is it not?

That’s right pickle is part of Python’s standard library, so there’s no need to install it via pip

[bazarkua]

And graphviz is installed via apt-get, I believe.

sudo apt-get install graphviz installs system level Graphviz executables actual engine toolkit (the dot tool, etc.) pip install graphviz installs the Python API that talks to installed system Graphviz engine (provides python classes and functions)

[bazarkua]

I believe Snakemake built in DAG commands emit a dot file and then shell out to the dot executable (the one that we install via apt), so we might not need to install it via pip

[ecwood]

Where is bmt used in the RTX-KG2 project code? I did a search both via GitHub's web browser UI and of the master branch code, like this:

sramsey-laptop:Desktop sramsey$ git clone [email protected]:RTXteam/RTX-KG2.git
Cloning into 'RTX-KG2'...
remote: Enumerating objects: 14409, done.
remote: Counting objects: 100% (1958/1958), done.
remote: Compressing objects: 100% (283/283), done.
remote: Total 14409 (delta 1743), reused 1676 (delta 1675), pack-reused 12451 (from 3)
Receiving objects: 100% (14409/14409), 4.28 MiB | 8.11 MiB/s, done.
Resolving deltas: 100% (9355/9355), done.
sramsey-laptop:Desktop sramsey$ cd RTX-KG2
sramsey-laptop:RTX-KG2 sramsey$ find . -type f -exec grep bmt {} /dev/null \;
sramsey-laptop:RTX-KG2 sramsey$

and did not get a hit.

Ok, cool. I am fine removing it then.

[saramsey]

@bazarkua: I think you can go ahead and close this issue, unless you wish to do additional checking.