This package contains three modules with classes useful for dealing with RxNorm, then a module each for UMLS and SNOMED handling.
Provides classes that deal with RxNorm. This is very much WiP!
Bases: object
A class for handling RxNorm in an SQLite database and performing a handful of RxNorm-related tasks.
Check if our database is in place and if not, import them. Will raise on errors!
RxNorm: (rxnorm.db) If missing prompt to use the rxnorm.sh script
Normalizes an NDC (National Drug Code) number.
The pseudo-code published by NIH (http://www.nlm.nih.gov/research/umls/rxnorm/NDC_Normalization_Code.rtf) first identifies the format (e.g. “6-3-2”) and then normalizes based on that finding. However since the normalized string is always 5-4-2, padded with leading zeroes and removing all dashes afterwards, this implementation goes a much simpler route.
NDCs that only contain one dash are treated as if they were missing the package specifier, so they get a “-00” appended before normalization.
| Parameters: | ndc (str) – The NDC to normalize as string |
|---|---|
| Returns: | A string with the normalized NDC, or None if the number couldn’t be normalized |
Bases: graphable.GraphableObject
Bases: graphable.GraphableRelation
Bases: object
Class for RxNorm lookup.
Execute and return the pointer of an SQLite execute() query.
Execute and return the result of fetchall() on a raw SQL query.
Execute and return the result of fetchone() on a raw SQL query.
Tries to reformat the VA drug class name so it’s suitable for display.
Returns a set of tuples containing the RXCUI and the actual relation for the desired relation, or all if the relation is not specified.
| Parameters: |
|
|---|---|
| Returns: | A set of tuples, where tuples are (rxcui, rela) |
Return a tuple with (str, tty, rxcui, rxaui) or - if “preferred” is False - a tuple with (preferred-name, list-of-tuples)
Return a string or HTML for the meaning of the given code. If preferred is True (the default), only one match will be returned, looking for specific TTY and using the “best” one. There is currently NO SUPPORT FOR preferred = False
Returns a set of TTYs for the given RXCUI.
Find the NDC from our NDC-cache-table for the given RXCUI.
Tries to find an RXCUI for the concept name.
Does this by performing a “starts with” against the STR column on RXNCONSO, then replaces any spaces with wildcards and finally chops off one word after the other until a match is found.
This works but is slow and far from perfect. RxNav’s approxMatch is definitely better, you can use rxcui_for_name_approx to get an RXCUI using that service.
| Parameters: |
|
|---|---|
| Returns: | The best matching rxcui, if any, as string |
Returns the best approxMatch RXCUI as found when using RxNav’s service against the provided name. Runs synchronously.
| Parameters: | name (str) – The name to get an RXCUI for |
|---|---|
| Returns: | The top ranked rxcui, if any, as string |
Find the RXCUI for the given NDC from our NDC-cache-table.
This method only does exact lookup for now, it should be extended to use normalized NDC formats.
| Parameters: | ndc (str) – The NDC to look up |
|---|---|
| Returns: | The matching RXCUI as string, or None |
Returns the top #nmax approximateTerm rxcuis as found when using RxNav’s service against the provided name. Runs synchronously.
| Parameters: |
|
|---|---|
| Returns: | The top ranked rxcuis, if any, as a list |
Returns a list of VA class names for a given RXCUI. EXPERIMENTAL.
A script used to create JSON documents from most RxNorm concepts and store them into a NoSQL database. This is very much WiP!
A useful script to help visualize relationships between RxNorm concepts, starting from a given RXCUI. Just run this script in your command line and follow the leader.
Module to deal with UMLS lexica.
Bases: object
A class for importing UMLS terminologies into an SQLite database.
Check if our database is in place and if not, prompts to import it. Will raise on errors!
UMLS: (umls.db) If missing prompt to use the umls.sh script
Bases: object
UMLS lookup
Return a list with triples that contain: - name - source - semantic type by looking it up in our “descriptions” database. The “preferred” settings has the effect that only names from SNOMED (SNOMEDCD) and the Metathesaurus (MTH) will be reported. A lookup in our “descriptions” table is much faster than combing through the full MRCONSO table.
| Returns: | A list of triples with (name, sab, sty) |
|---|
Tries to find a good concept code for the given concept name.
Uses our indexed descriptions table.
| Returns: | A list of triples with (cui, sab, sty) |
|---|
Return a string (an empty string if the cui is null or not found) by looking it up in our “descriptions” database. The “preferred” settings has the effect that only names from SNOMED (SNOMEDCD) and the Metathesaurus (MTH) will be reported. A lookup in our “descriptions” table is much faster than combing through the full MRCONSO table.
Module to deal with the SNOMED terminology.
Bases: object
A class for importing UMLS terminologies into an SQLite database.
Check if our database is in place and if not, prompts to create it. Will raise on errors!
SNOMED: (snomed.db) Read SNOMED CT from tab-separated files and create an SQLite database.
Allows us to set hooks after tables have been imported
Import SNOMED CSV into our SQLite database. The SNOMED CSV files can be parsed by Python’s CSV parser with the “excel-tab” flavor.
Returns the insert query needed for the given table
Creates the SQLite tables we need, not the tables we deserve. Does nothing if the tables/indexes already exist
Provides classes that can be used to create an interdependency graph.
Bases: object
Bases: object
Announce the receiver to the context.
Subclasses MUST NOT announce other graphable objects they are holding on to here but they MUST announce them in “deliver_to” if appropriate.
dot_context The context to announce to
announcing the receiver
Call the context’s “deliver” method.
This method is guaranteed to only be called once per context. Hence subclasses that hold on to other graphable objects MUST ANNOUNCE those instances here (but NOT deliver them) but ONLY IF “is_leaf” is not True. - dot_context The context to deliver to - is_leaf If True means the receiver is intended to be a leaf object
Our SQLite connection class.
Bases: object
A simple CSV to SQLite importer class.
Expects a CSV file with a header row, will create a table reflecting the header row and import all rows.
Bases: object
SQLite access
Executes a CREATE TABLE IF NOT EXISTS query with the given structure. Input is NOT sanitized, watch it!
Executes an SQL command and returns the cursor.execute, which can be used as an iterator. Supply the params as tuple, i.e. (param,) and (param1, param2, ...)
Executes an SQL command (should be INSERT OR REPLACE) and returns the last row id, 0 on failure.
Returns the first row returned by executing the command
Executes an SQL command (should be UPDATE) and returns the number of affected rows.
Use this to get SQLite instances for a given database. Avoids creating multiple instances for the same database.
We keep instances around per thread per database, maybe there should be a way to turn this off. However, here we always release instances for threads that are no longer alive. If this is better than just always creating a new instance should be tested.
Returns whether the given table exists.