db attribute error from 'if datacache.db.db_table_exists(db, 'ensembl'):' in database.py

[Xerez13]

I recently installed and ran pyensembl on my machine. The first time I called the module in an interactive session everything worked great.

from pyensembl import EnsemblRelease
data = EnsemblRelease(77)
gene_name = data.gene_names_at_locus(contig=11,position=101567059)
print gene_name

However, in any subsequent calls, typing in the same code as above, I received the following error:

Traceback (most recent call last): File "", line 1, in File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/ensembl_release.py", line 1 65, in gene_names_at_locus File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 226, in distinct_column_values_at_locus File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 150, in column_values_at_locus File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 120, in column_exists File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 116, in columns File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 111, in connection File "build/bdist.macosx-10.5-x86_64/egg/pyensembl/database.py", line 99, in _connect_or_create_database AttributeError: 'module' object has no attribute 'db'

When I went to the _connect_or_create_database function within database.py I found the following line was throwing the error:

if datacache.db.db_table_exists(db, 'ensembl'):

When I looked at the module datacache I found neither the db nor db_table_exists function:

dir(datacache) ['Cache', 'builtins', 'doc', 'file', 'loader', 'name', 'package', 'path', 'build_local_filename', 'build_path', 'cache', 'clear_cache', 'common', 'connect_if_correct_version', 'database', 'database_helpers', 'database_table', 'database_types', 'db_from_dataframe', 'db_from_dataframes', 'db_from_dataframes_with_absolute_path', 'download', 'ensure_dir', 'fasta', 'fetch_and_transform', 'fetch_csv_dataframe', 'fetch_csv_db', 'fetch_fasta_db', 'fetch_fasta_dict', 'fetch_file', 'get_data_dir']

My version of datacache was 0.4.16

The line which throws the error appears to be calling a functionality which is not present in datacache. The code initially works fine because if db_path does not exist the function directs to _create_database() which contains functions present in datacache. However, in subsequent calls the function attempts to execute the troublesome line and throws the error. I was able to get around this error by commenting out the line in question along with the return statement below it. I inserted a return statement below the db assignment. A copy of my adjustment is shown below.

    def _connect_or_create_database(self):
        """
        If database already exists, open a connection.
        Otherwise, create it.
        """
        db_path = self.local_db_path()
        if exists(db_path):
            db = sqlite3.connect(db_path)
            return db #edit
            # maybe file got created but not filled
            #if datacache.db.db_table_exists(db, 'ensembl'): #edit
            #    return db #edit
        return self._create_database()

I just thought I would raise this issue since this tool is quite cool and has great functionality, but this error might frustrate people if they aren't willing to investigate it.

[iskandr]

Hey @Xerez13,

Thanks for this detailed report. I'm not, however, able to reproduce the problem! In the most recent master branch of pyensembl, line 99 of database.py is:

            ['transcript_name'],

and the expression datacache.db.db_table_exists doesn't occur anywhere in the file.

Any chance you're using an older version of pyensembl with a more recent datacache?

[Xerez13]

Hey @iskandr ,

I just checked and I am using pyensembl-0.5.4 which appears to be an older version of the package. My datacache version is 0.4.16. It is likely this error comes about exactly for the reason you suggest.

[rohandavidg]

i had the same issue..datacache-0.4.16-py2.7.egg, pyensembl-0.5.4-py2.7.egg

[iskandr]

Hey Rohan,

Can you try upgrading your version of PyEnsembl to the more recent 0.8.4?

On Sun, Feb 14, 2016 at 5:58 PM, Rohan Gnanaolivu notifications@github.com wrote:

i had the same issue..datacache-0.4.16-py2.7.egg, pyensembl-0.5.4-py2.7.egg

— Reply to this email directly or view it on GitHub https://github.com/hammerlab/pyensembl/issues/133#issuecomment-183998439 .

[iskandr]

I'm going to close since I suspect this error only occurs when an older PyEnsembl is installed with a newer datacache.