GREP on cloud
A GNU grep-like tool that works on Amazon AWS. It is called "fdsm", which stands for file directory search and manipulation.
Program Execution
fdsm [-b] [-f] [-s <path/to/statementsfile>] [-i <instance>] [<directory/to/instance>]
The program options are as follows:
-b: Traverse the directory in BFS manner. (default: DFS)
-f: Use full path name. (default: run on current path)
-s: Get statements from a file (given path/to/statementsfile). (default: get statements from stdin)
-i: Run on Amazon Cloud machine with given instance id. (default: on host machine)
directory: Choose in which directory the code will run on the cloud
Syntax
Statements file has the following syntax:
<fdsm expression> => <Unix/Python commands separated by semicolon>;
...
For example:
start => {count = 0} ;
/^A/ && file && d/01-04-2002/b => echo $MATCHED ; rm $MATCHED ;
(/^core$/ || /.+\.o$/ ) && file => rm $MATCHED ;
/^.*\.py$/ && file => {count = count+1\nprint $MATCHED} ;
/junk/ && directory => rm $MATCHED ;
finish => {print count} ;
The expected behavior of the script above is:
count = 0
$MATCHED = [files which are created before 01.04.2002 and have names starting with A]
print $MATCHED
remove $MATCHED
$MATCHED = [files with name "core" or extension .o]
remove $MATCHED
$MATCHED = [.py files]
count = count + 1
remove $MATCHED
$MATCHED = [directories with name "junk"]
remove $MATCHED
print count
The statements on the right can be either python or unix statements. Python statements are in curly brackets ({}).
The statements on the left are in fdsm syntax which are patterns that will match file properties. The patterns are as follows:
| syntax | meaning |
|---|---|
| start, finish | beginning and the end of the code block |
| file, directory | if a path is a directory or a file |
| /<regex-statement>/ | item name matches |
| c/<regex-statement>/ | there is a match in file content |
| o/<regex-statement>/ | owner of the item matches |
| p/<regex-statement>/ | permission type of the item matches |
| d/<regex-statement>/[b/a] | creation date of the item matches (options: before/after) |
| s/<regex-statement>/[l] | size of the item matches (options: larger) |
| readable, writeable, executable | checks these permissions on the item |