X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/e32be386f2378807a79da5810414982769aa046c..23053a2d35694a9e8a7169c493620c3a33bd78f2:/crunch_scripts/run-command diff --git a/crunch_scripts/run-command b/crunch_scripts/run-command index 916e6b2d7f..3b58186171 100755 --- a/crunch_scripts/run-command +++ b/crunch_scripts/run-command @@ -1,7 +1,12 @@ #!/usr/bin/env python import logging -logging.basicConfig(level=logging.INFO, format="run-command: %(message)s") + +logger = logging.getLogger('run-command') +log_handler = logging.StreamHandler() +log_handler.setFormatter(logging.Formatter("run-command: %(message)s")) +logger.addHandler(log_handler) +logger.setLevel(logging.INFO) import arvados import re @@ -86,6 +91,10 @@ def expand_item(p, c): params[var] = i r.extend(expand_list(params, c["command"])) return r + if "list" in c and "index" in c: + var = c["list"] + items = get_items(p, p[var]) + return items[int(c["index"])] elif isinstance(c, list): return expand_list(p, c) elif isinstance(c, str) or isinstance(c, unicode): @@ -96,7 +105,34 @@ def expand_item(p, c): def expand_list(p, l): return [exp for arg in l for exp in expand_item(p, arg)] +def add_to_group(gr, match): + m = ('^_^').join(match.groups()) + if m not in gr: + gr[m] = [] + gr[m].append(match.group(0)) + def get_items(p, value): + if isinstance(value, dict): + if "filter" in value and "regex" in value: + pattern = re.compile(value["regex"]) + items = get_items(p, value["group"]) + return [i for i in items if pattern.match(i)] + + if "group" in value and "regex" in value: + pattern = re.compile(value["regex"]) + items = get_items(p, value["group"]) + groups = {} + for i in items: + p = pattern.match(i) + if p: + add_to_group(groups, p) + return [groups[k] for k in groups] + + if "extract" in value and "regex" in value: + pattern = re.compile(value["regex"]) + items = get_items(p, value["group"]) + return [p.groups() for i in items if p = pattern.match(i)] + if isinstance(value, list): return expand_list(p, value) @@ -108,7 +144,7 @@ def get_items(p, value): items = ["$(dir %s/%s/)" % (prefix, l) for l in os.listdir(fn)] elif stat.S_ISREG(mode): with open(fn) as f: - items = [line for line in f] + items = [line.rstrip("\r\n") for line in f] return items else: return None @@ -119,33 +155,50 @@ stdinname = None stdinfile = None rcode = 1 +def recursive_foreach(params, fvars): + var = fvars[0] + fvars = fvars[1:] + items = get_items(params, params[var]) + logger.info("parallelizing on %s with items %s" % (var, items)) + if items is not None: + for i in items: + params = copy.copy(params) + params[var] = i + if len(fvars) > 0: + recursive_foreach(params, fvars) + else: + arvados.api().job_tasks().create(body={ + 'job_uuid': arvados.current_job()['uuid'], + 'created_by_job_task_uuid': arvados.current_task()['uuid'], + 'sequence': 1, + 'parameters': params + } + ).execute() + else: + logger.error("parameter %s with value %s in task.foreach yielded no items" % (var, params[var])) + sys.exit(1) + try: if "task.foreach" in jobp: if arvados.current_task()['sequence'] == 0: - var = jobp["task.foreach"] - items = get_items(jobp, jobp[var]) - logging.info("parallelizing on %s with items %s" % (var, items)) - if items is not None: - for i in items: - params = copy.copy(jobp) - params[var] = i - arvados.api().job_tasks().create(body={ - 'job_uuid': arvados.current_job()['uuid'], - 'created_by_job_task_uuid': arvados.current_task()['uuid'], - 'sequence': 1, - 'parameters': params - } - ).execute() - if "task.vwd" in jobp: - # Base vwd collection will be merged with output fragments from - # the other tasks by crunch. - arvados.current_task().set_output(subst.do_substitution(jobp, jobp["task.vwd"])) - else: - arvados.current_task().set_output(None) - sys.exit(0) - else: + # This is the first task to start the other tasks and exit + fvars = jobp["task.foreach"] + if isinstance(fvars, basestring): + fvars = [fvars] + if not isinstance(fvars, list) or len(fvars) == 0: + logger.error("value of task.foreach must be a string or non-empty list") sys.exit(1) + recursive_foreach(jobp, jobp["task.foreach"]) + if "task.vwd" in jobp: + # Set output of the first task to the base vwd collection so it + # will be merged with output fragments from the other tasks by + # crunch. + arvados.current_task().set_output(subst.do_substitution(jobp, jobp["task.vwd"])) + else: + arvados.current_task().set_output(None) + sys.exit(0) else: + # This is the only task so taskp/jobp are the same taskp = jobp if "task.vwd" in taskp: @@ -165,16 +218,16 @@ try: stdoutname = subst.do_substitution(taskp, taskp["task.stdout"]) stdoutfile = open(stdoutname, "wb") - logging.info("{}{}{}".format(' '.join(cmd), (" < " + stdinname) if stdinname is not None else "", (" > " + stdoutname) if stdoutname is not None else "")) + logger.info("{}{}{}".format(' '.join(cmd), (" < " + stdinname) if stdinname is not None else "", (" > " + stdoutname) if stdoutname is not None else "")) except subst.SubstitutionError as e: - logging.error(str(e)) - logging.error("task parameters was:") - logging.error(pprint.pformat(taskp)) + logger.error(str(e)) + logger.error("task parameters were:") + logger.error(pprint.pformat(taskp)) sys.exit(1) except Exception as e: - logging.exception("caught exception") - logging.error("task parameters was:") - logging.error(pprint.pformat(taskp)) + logger.exception("caught exception") + logger.error("task parameters were:") + logger.error(pprint.pformat(taskp)) sys.exit(1) try: @@ -190,13 +243,13 @@ try: rcode = sp.wait() if sig.sig is not None: - logging.critical("terminating on signal %s" % sig.sig) + logger.critical("terminating on signal %s" % sig.sig) sys.exit(2) else: - logging.info("completed with exit code %i (%s)" % (rcode, "success" if rcode == 0 else "failed")) + logger.info("completed with exit code %i (%s)" % (rcode, "success" if rcode == 0 else "failed")) except Exception as e: - logging.exception("caught exception") + logger.exception("caught exception") # restore default signal handlers. signal.signal(signal.SIGINT, signal.SIG_DFL) @@ -206,11 +259,11 @@ signal.signal(signal.SIGQUIT, signal.SIG_DFL) for l in links: os.unlink(l) -logging.info("the following output files will be saved to keep:") +logger.info("the following output files will be saved to keep:") subprocess.call(["find", ".", "-type", "f", "-printf", "run-command: %12.12s %h/%f\\n"], stdout=sys.stderr) -logging.info("start writing output to keep") +logger.info("start writing output to keep") if "task.vwd" in taskp: if "task.foreach" in jobp: @@ -220,7 +273,7 @@ if "task.vwd" in taskp: # Just a single task, so do merge with the original collection outcollection = vwd.checkin(subst.do_substitution(taskp, taskp["task.vwd"]), outdir, merge=True).manifest_text() else: - outcollection = robust_put.upload(outdir) + outcollection = robust_put.upload(outdir, logger) api.job_tasks().update(uuid=arvados.current_task()['uuid'], body={