This package maps changes in drug label to patent claims using the Orange Book Table's of Exclusivity.
Running this package without any optional argument will calculate and store diffs between adjacent labels (by date) for each drug as defined by NDA number(s) into the label collection of the MongoDB database. If the diff is an addition with alphanumeric characters, it will be mapped to patent claims from the patents collection of the MongoDB database.
Labels that are already processed are stored in resources/processed_log/processed_id_diff.csv and resources/processed_log/processed_id_similarity.csv. These labels will not be re-processed unless optional argument -r is set. Running optional arguments other than -r will not additionally run the diffing steps or diffs-to-patent-claims mapping unless those flags are set.
Running this package with the -since flag will cause this package to update the diff and similarity for all labels on or since the date passed to the -since flag.
Regarding the label collection, this package adds the the following information to labels stored in a MongoDB database:
previous_label_published_date: null,
previous_label_spl_id: null,
previous_label_spl_version: null,
next_label_published_date: '2019-07-24',
next_label_spl_id: '0c01a1fd-afc2-4c3e-9d5a-5000fa5a9339',
next_label_spl_version: '2',
diff_against_previous_label: [
{
name: '1 INDICATIONS AND USAGE',
text: [
[
1,
'DRIZALMA Sprinkle is indicated for the treatment of:\n•Major Depressive Disorder in adults [see Clinical Studies (14.1)]\n•Generalized Anxiety Disorder in adults and pediatric patients 7 years to 17 years old [see Clinical Studies (14.2)]\n•Diabetic Peripheral Neuropathy in adults [see Clinical Studies (14.3)]\nChronic Musculoskeletal Pain in adults [see Clinical Studies (14.4)]',
'0',
{
expanded_content: 'DRIZALMA Sprinkle is indicated for the treatment of:\n•Major Depressive Disorder in adults [see Clinical Studies (14.1)]\n•Generalized Anxiety Disorder in adults and pediatric patients 7 years to 17 years old [see Clinical Studies (14.2)]\n•Diabetic Peripheral Neuropathy in adults [see Clinical Studies (14.3)]\nChronic Musculoskeletal Pain in adults [see Clinical Studies (14.4)]',
scores: [
{
patent_number: '9839626',
claim_number: 16,
parent_claim_numbers: [],
score: 0.4180474579334259
},
...
]
},
...
],
...
]
},
...
]
additions: {
'0': {
expanded_content: 'DRIZALMA Sprinkle is indicated for the treatment of:\n•Major Depressive Disorder in adults [see Clinical Studies (14.1)]\n•Generalized Anxiety Disorder in adults and pediatric patients 7 years to 17 years old [see Clinical Studies (14.2)]\n•Diabetic Peripheral Neuropathy in adults [see Clinical Studies (14.3)]\nChronic Musculoskeletal Pain in adults [see Clinical Studies (14.4)]',
scores: [
{
patent_number: '9839626',
claim_number: 16,
parent_claim_numbers: [],
score: 0.4180474579334259
},
...
]
},
...
}
nda_to_patent: [
{
application_number: '212516',
patents: [
'10413525',
'9839626'
]
}
]
diff_against_previous_label includes all changes versus the previous label for the same set-id.
Note that the first index of each element of text within diff_against_previous_label indicates whether diff_against_previous_label is an addition 1, subtraction -1, or no change 0 relative to the previous labeled with the same set-id.
The second index of each element of text within diff_against_previous_label indicates the actual text in question.
The third index of each element of text within diff_against_previous_label refers to the key in additions
The fourth index of each element of text within diff_against_previous_label copies the value of additions for the key in the third index of each element of text within diff_against_previous_label. This is a request from the front end developers to ease development.
additions stores all additions by a key, which is the value in the third index of each element of text within diff_against_previous_label.
expanded_context refers to the text within the second index of each element of text within diff_against_previous_label, when the first index of each element of text within diff_against_previous_label is 1, but is expanded in some cases to include up to the start and end of the sentence surrounding the second index of each element of text within diff_against_previous_label. For example, if the addition were just one word, expanded context will include the entire sentence which has that word.
scores list the related claim for each addition in order from highest to lowest scoring claim for all related patent_numbers. The related patent_number are determined by an Orange Book lookup.
nda_to_patent lists all related patent for each NDA number.
The connection info for the Mongo DB instance is set in the .env file. This should work for a standard MongoDB set up on localhost. If using a different set of DB configs, this file must be updated.
A sample docker set up for mongo, that maps to localhost:27017 can be found here. This set up also includes the Mongo Express viewer.
To load the a stale-state json collection files (current as of 2021-05-12) into the docker container, after running docker-compose up per the docker and MongoDB setup here, find the docker container using docker ps -a. Assuming that docker ps -a indicates that mongo_local is the name of the docker container that is running MongoDB, then confirm that MONGODB_HOST and MONGODB_PORT from .env matches the address and port underneath PORTS of docker ps -a for mongo_local.
Then uncompress both database_latest.tar.gz and processed_log.tar.gz in the resources/ folder. This should create resources/database_latest and resources/processed_log folders.
Then you may use either of the following two sets of commands to loads the latest collections into the databases:
docker cp resources/database_latest/labels.json mongo_local:.
docker cp resources/database_latest/labelmap.json mongo_local:.
docker cp resources/database_latest/patents.json mongo_local:.
docker cp resources/database_latest/orangebook.json mongo_local:.
docker exec -it mongo_local sh
mongoimport --db latest --collection labels --file labels.json
mongoimport --db latest --collection labelmap --file labelmap.json
mongoimport --db latest --collection patents --file patents.json
mongoimport --db latest --collection orangebook --file orangebook.json
or
Change MONGODB_NAME to latest in .env. Then run python3 main.py -rip -ril -rilm -rio once the Install instructions below are completed. This will load the resources/database_pharmadb/ collections into MongoDB. (Warning: python3 main.py -rip -ril -rilm -rio is for the import of collections during testing and development period, and should not be used for production! In production, the databases are already populated by scripts from other repositories. In which case, this module merely performs the diffing and similarity comparisons on data that is already stored on MongoDB.)
Requires a minimum python version of 3.6 to run.
To install:
pip3 install -r requirements.txt
or
pipenv install
If pipenv is used all subsequent python3 commands should be replaced with pipenv run python.
To run sequential label diff and similarity comparisons of diffs to patent claims (from scratch):
python3 main.py -r
If -r is omitted, labels that have already been processed and stored in resources/processed_log with not be reprocessed.
Alternatively, to run sequential label diff and similarity comparisons of diffs to patent claims for all NDAs based on update date in the orangebook MongoDB collection:
python3 main.py -since <YYYY-MM-DD>
To download latest Orange Book:
python3 main.py -ob
To output a list of all NDA numbers from the Orange Book:
python3 main.py -an <filename>
To output a list of all patents from the Orange Book:
python3 main.py -ap <filename>
To output a list of all patents from the Orange Book as a json file:
python3 main.py -apj <filename>
To output a list of missing NDA from the database that are in the Orange Book:
python3 main.py -mn <filename>
To output a list of missing patents from the database that are in the Orange Book:
python3 main.py -mp <filename>
To output a list of missing patents from the database that are in the Orange Book as a json file:
python3 main.py -mpj <filename>
These outputs will require confirmation if the file already exists, so use the following to output all files into the asset/ directory:
yes | python3 main.py -ob -an -ap -apj -mn -mp -mpj
To read help:
python3 main.py -h
To output all addition and patent claim set from the database excluding all scores into folders.
python3 main.py -db2file <folder_name>
If <folder_name> is not set, file is stored in the analysis folder.
Alternatively, a compressed version of the export with stale data is located at analysis/db2file.tar.gz
To output all addition and patent claim set from the database including all scores into a single .csv file.
python3 main.py -db2csv <filename>
The <filename> is optional. If not set folder is stored in /resources/hosted_folder/db2csv.zip.
A compressed version of the export with stale data is located at resources/hosted_folder/db2csv.zip
The zipped file can be hosted on a webserver that can be started with:
sudo nohup python3 server.py &
Unit tests are run with:
python3 -m unittest
It is recommended to use the Black Code Formatter which can be installed as a plugin for most IDEs. pyproject.toml holds the formatter settings.