- Adding links between trials and interventions in addition to trials and conditions.
- conditions and interventions are fetched from database (instead of files). Corresponding edges b/w trials and conditions, trials and interventions are added to RDF. For example :
<https://clinicaltrials.gov/ct2/show/NCT00093782> <Condition> <http://id.nlm.nih.gov/mesh/2022/T000687> .
<https://clinicaltrials.gov/ct2/show/NCT00093782> <Intervention> <http://id.nlm.nih.gov/mesh/2022/T538652> .
- All global trial's - 755450 - are added to RDF. For example :
<https://clinicaltrials.gov/ct2/show/NCT00172328> <TrialId> "NCT00172328" .
<https://www.who.int/clinical-trials-registry-platform/CTRI/2021/05/033487> <TrialId> "CTRI/2021/05/033487" .
- Starting with a fresh model for final RDF instead of MeSH RDF. MeSH ids that are not linked to any trial are not considered. This reduces the graph size considerably.
- Trial records are fetched from ICTRP's weekly + periodic full export and AACT's daily + monthly full snapshot.
- Trials are written down to a file (will be used later) : vaidhyamegha_clinical_trials.csv.gz
$ cat data/open_knowledge_graph_on_clinical_trials/vaidhyamegha_clinical_trials.csv | sort -u | wc -l
755450
- Download the RDF : vaidhyamegha_open_kg_clinical_trials_v0.3.nt.gz