What is left to do:

I would appreciated feedback and help on all these issues and where next to take ruffus.

Please write to me ( ruffus_lib at llew.org.uk) or join the project.

Some of these proposals are well-fleshed-out:

• Clean up
• (Plug-in) File Dependency Checking via MD5 or Databases

Others require some more user feedback about semantics:

• Combining output from multiple jobs automatically
• Harvesting return values from jobs
• Linking the output from one task as the input to the next automatically

Some issues are do-able but difficult and I don’t have the experience:

• Run jobs on remote (clustered) processes via SGE/Hadoop

Linking the output from one task as the input to the next automatically

For each task, we would have to log all the output files from each job, and use that as the input into the next stage.

This is, after all, what make does.

Except that we would be creating the dependency tree on the fly.

Syntactically, it would make things a bit easier for the user, with minor additions to the @file and @file_re syntax.

Clean up

The plan is to store the files and directories created via a standard interface.

The placeholders for this are a function call register_cleanup.

Jobs can specify the files they created and which need to be deleted by returning a list of file names from the job function.

So:

raise Exception = Error

return False = halt pipeline now

return string / list of strings = cleanup files/directories later

return anything else = ignored

The cleanup file/directory store interface can be connected to a text file or a database.

The cleanup function would look like this:

pipeline_cleanup(cleanup_log("../cleanup.log"), [instance ="october19th" ])
pipeline_cleanup(cleanup_msql_db("user", "password", "hash_record_table"))

The parameters for where and how to store the list of created files could be similarly passed to pipeline_run as an extra parameter:

pipeline_run(cleanup_log("../cleanup.log"), [instance ="october19th" ])
pipeline_run(cleanup_msql_db("user", "password", "hash_record_table"))

where cleanup_log and cleanup_msql_db are classes which have functions for

1. storing file
2. retrieving file
3. clearing entries

• Files would be deleted in reverse order, and directories after files.

• By default, only empty directories would be removed.

  But this could be changed with a --forced_remove_dir option

• An --remove_empty_parent_directories option would be supported by os.removedirs(path).

(Plug-in) File Dependency Checking via MD5 or Databases

So that MD5 / a database can be used instead of coarse-grained file modification times.

As always, the design is a compromise between flexibility and easy of use.

The user can already write their own file dependency checking function and supply this:

@check_if_uptodate(check_md5_func)
@files(io_files)
def task_func (input_file, output_file):
    pass

The question is can we

1. supply a check_md5() function
2. allow the whole pipeline to use this.

Most probably we need an extra parameter somewhere:

pipeline_run(md5_hash_database = "current/location/files.md5")

There is prior art on this in scons.

If we use a custom object/function, can we use orthogonal syntax for

1. disk modifications times,
2. md5 hashes saved to a file,
3. md5 hashes saved to a database?

pipeline_run(file_up_to_date_lookup = md5_hash_file("current/location/files.md5"))
pipeline_run(file_up_to_date_lookup = mysql_hash_store("user", "password", "hash_record_table"))

where md5_hash_file and mysql_hash_store are objects which have get/set functions for looking up modification times from file names.

Of course that allows you to fake the whole process and not even use real files...

SQL hooks

See above.

I have no experience with systems which link to SQL. What would people want from such a feature?

Ian Holmes?

Combining files

A common operation (See intermediate example and complicated example) seems to be combining multiple input files into one or more output files (i.e. a “many->1” operation).

A the moment, this can be easily done by writing a simple parameter generating function with a glob (see this code.)

However, we could directly support this in ruffus in two ways, both leveraging @files_re.

The idea is that input files to be grouped together should share similar traits identified by either a regular expression or a function.

Suppose we had the following files:

cows.mammals.animal
horses.mammals.animal
sheep.mammals.animal

snake.reptile.animal
lizard.reptile.animal
crocodile.reptile.animal

pufferfish.fish.animal

and we wanted to end up with three different resulting output:

cow.mammals.animal
horse.mammals.animal
sheep.mammals.animal
    -> mammals.results

snake.reptile.animal
lizard.reptile.animal
crocodile.reptile.animal
    -> reptile.results

pufferfish.fish.animal
    -> fish.results

We could either add a “combining” regex parameter:

# \1 = species [cow, horse]
# \2 = phylogenetics group [mammals, reptile, fish]
@files_re(  "*.animal",
            r"(.+)\.(.+)\.animal",
            combining(r"\2"),           # regular expression showing that
                                        # "mammals/reptile" etc. groups files
            r"\1.\2.animal",            # input file
            r"\2.results")              # output file(s)
def sort_animals_into_groups(species_file, result_file):
    " ... more code here"

Or we could allow a grouping function like the key parameter in itertools.groupby:

@files_re(  "*.animal",
            r"(.+)\.(.+)\.animal",
            combining_func(lambda x: x.split(".")[1]),
            r"\1.\2.animal",            # input file
            r"\2.results")              # output file(s)
def sort_animals_into_groups(species_file, result_file):
    " ... more code here"

In both cases, the extra parameter would be wrapped by a “tagging” class (combining and combining_func) for clarity.

Is this too much extra complexity for ruffus or @files_re to support?

Is the syntactic convenience worthwhile?

Return values

Is it a good idea to allow jobs to pass back calculated values?

This requires trivial modifications to run_pooled_job_without_exceptions

The user function cannot return False. That is reserved for signalling breaks.

The most useful thing would be to associate job parameters with results.

What should be the syntax for getting the results back?

Run jobs on remote (clustered) processes via SGE/Hadoop

Can we run jobs on remote processes using SGE / Hadoop?

Can we abstract all job management using drmaa?

Python examples at http://gridengine.sunsource.net/howto/drmaa_python.html

SGE

Look at Qmake execution model:

1) SGE nodes are taken over completely

See last example in multiprocessing for creating a distributed queue.

We would use qrsh instead of ssh. The size of the pool would be the (maximum) number of jobs

Advantages:

• Simple to implement
• Efficient

Disadvantages:

• Other users might not appreciate having python jobs taking over the nodes for a protracted length of time
• We would not be able to use SGE to view / manage jobs

2) Start a qrsh per job

Advantages:

• jobs look like any other SGE task

Disadvantages:

• Slower. Overheads might be high.
• We might have to create a new pool per task
• If we maintain an empty pool, and then dynamically attach processes, the code might be difficult to write (may not fit into the multiprocessing way of doing things / race-conditions etc.)

Hadoop

Can anyone help me with this / have any experience?