Annotation of Adverse Drug Reactions in Patients’Weblogs

[Japanese]

We only distribute metadata of annotations as json files. Please follow the steps below to reconstruct the annotations with full contents. Basically, you only need to run run.sh after setting appropriate arguments.

1. Download the medatada json file.
2. Create an empty folder and put the medatada json file.
3. Edit arguments of run.sh.
   • Set the folder name created in step-2 to --folder_path.
   • Set the json file name to --infile_name.
   • Set an output file name to --out_file. (* extension must be .json.)
4. Run sh run.sh.
   • It downloads all the weblog articles over the internet through http using request of python module.
   • It creates ./folder_path/html/ and ./folder_path/article/ (You may delete these later if not needed).
   • There is a one second interval between request. Depending on the internet connection, this step may take time. Please be patient.
5. The json file with full text content will be reconstructed and saved.
   • The location was specified by --folder_path.
   • The output file name was specified by --out_file.

You may try and test this script using a sample file. Please simply run sh run.sh to generate ./trial/annotation.json file from./trial/trial.json file.