Skip to content
This repository has been archived by the owner on Sep 17, 2019. It is now read-only.

pekrau/adhoc

Repository files navigation

Adhoc: A simple web application for task execution

This web application is an interface to various third-party software, which are made available as a set of tools for executing specific tasks.

Available tools

The set of BLAST programs for searching biomolecule sequence databases is available. The list of BLAST databases includes some public standard data sets. Depending on the teams that your account is a member of, some private databases may also be available.

Other tools can be implemented as the need arises. Contact the administrator.

Databases

The databases for the tools are managed by the administrator. It is currently not possible for a user to upload a database through the web interface. The access to the databases is controlled by the team memberships of a user account.

Contact the administrator if you wish to make additional databases available, or if your memberships should be changed. Instructions for the administrator are available below.

Usage

Each software executable is represented as a tool. The tool page contains input fields to define a task to be executed by the tool. The user provides the input and submits the task, which is then executed by the server in the background.

Each task is identified by a IUI (Instance Unique Identifier), which is a string of 32 characters, visible in the URL of the task. The user can bookmark this URL to view the task and check its current status at any time. When the task has finished, the output can be viewed or downloaded. A task is persistent; it is stored on the server until explicitly deleted by the user.

In the current implementation, the execution is simply done in the background on the server. Other implementations may use a queueing system and/or execute the tasks on other back-end machines.

Login

Usually, a user logs in to an account before using the system. Any tasks created will then be private, and some private databases may become available, depending on the team memberships of the user's account.

It is possible to use the Adhoc system without logging in, in which case the built-in account 'anonymous' is used. All its tasks are publicly available, and only public databases are available to it.

Accounts and teams are administered in the WhoYou system. Contact the administrator of the server to obtain your own account.

RESTful interface

The Adhoc web interface is RESTful, meaning that scripts can easily be written to access the system from other machines.

Implementation

The system is written in Python 2.6. The following source code packages are needed:.

The Sqlite3 database system is used as storage back-end in the current implementation. It is included in the standard Python distribution.

An example installation can be viewed at http://tools.scilifelab.se/adhoc.

How to include a new database

The databases for the BLAST tool are located in /var/local/adhoc/db/blast

The file log.txt in that directory contains a record of what has been added and how it was done.

  1. Move the FASTA file to the blast directory.
  2. Run the makeblastdb command on the FASTA file. See the log.txt file for examples for the correct type of sequence database.
  3. Record the command and its output to the log.txt file.
  4. Add a record to the database configuration file "protein_databases.json" or "nucleotide_database.json". Specify "size" as 1 for the time being. The "number" parameter is the number of sequences, which is obtained from the output of the makeblastdb command.
  5. Run a test using some common query sequence. The total number of letters (bases or residues) should be entered (as an integer without commas) into the database configuration file.
  6. Adjust ownership and protection of all files created.
  7. Make copies of the original database files to /var/local/db-backup/blast directory, since only that directory is backed up.

Note that the teams are defined in the WhoYou system.

About

Task-oriented web interface system to various bioinformatics tools.

Resources

Stars

Watchers

Forks

Packages

No packages published