1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
|
# TODO: Add /jobs/<id>/data.zip to fetch all files at once in a zip file format
# TODO: Store completed work in /path/to/store/<id>.zip
import os
import logging
import json
import pickle
import flask
from flask import redirect, url_for, flash
from flask import send_from_directory
from werkzeug.utils import secure_filename
from . import store
app = flask.Flask(__name__)
# ==== File upload specialization ===
# By uploading files into a temporary file provided by store, we
# can then move the files directly into place on the store rather
# than copy them. This gives us reduced storage, reduced memory
# and reduced CPU.
class Request(flask.Request):
# Upload directly into temporary files.
def _get_file_stream(self, total_content_length, content_type,
filename=None, content_length=None):
#print "returning named temporary file for",filename
return store.tempfile()
app.request_class = Request
# ==== Format download specialization ===
def _format_response(response, format='json', template=None):
"""
Return response as a particular format.
"""
#print "response",response
if format == 'html':
if template is None: flask.abort(400)
return flask.render_template(template, **response)
elif format == 'json':
return flask.jsonify(**dict((str(k),v) for k,v in response.items()))
elif format == 'pickle':
return pickle.dumps(response)
else:
flask.abort(400) # Bad request
@app.route('/jobs.<format>', methods=['GET'])
def list_jobs(format='json'):
"""
GET /jobs.<format>
Return a list of all job ids.
"""
response = dict(jobs=SCHEDULER.jobs())
return _format_response(response, format, template='list_jobs.html')
@app.route('/jobs/<any(u"pending",u"active",u"error",u"complete"):status>.<format>',
methods=['GET'])
def filter_jobs(status, format='json'):
"""
GET /jobs/<pending|active|error|complete>.<format>
Return all jobs with a particular status.
"""
response = dict(jobs=SCHEDULER.jobs(status=str(status).upper()))
return _format_response(response, format, template='list_jobs.html')
@app.route('/jobs.<format>', methods=['POST'])
def create_job(format='json'):
"""
POST /jobs.<format>
Schedule a new job, return the job record.
The POST data should contain::
{
notify: "<user@email or @twitterid>",
service: "<name of service>",
version: "<service version>",
name: "<request name>",
data: "<service data>",
...
}
The response contains::
{
id: <job id>,
job: <job details>
}
Job details is simply a copy of the original request.
"""
request = flask.request.json #@UndefinedVariable in request proxy
if request is None: flask.abort(415) # Unsupported media
id = SCHEDULER.submit(request, origin=flask.request.remote_addr) #@UndefinedVariable in request proxy
flash('Job %s scheduled' % id)
response = {'id': id, 'job': SCHEDULER.info(id)}
#return redirect(url_for('show_job', id=id, format=format))
return _format_response(response, format=format, template='show_job.html')
@app.route('/jobs/<int:id>.<format>', methods=['GET'])
def show_job(id, format='json'):
"""
GET /jobs/<id>.<format>
Get job record by id.
The response contains::
{
id: <job id>,
job: <job details>
}
Job details is simply a copy of the original request.
"""
response = {'id': id, 'job': SCHEDULER.info(id)}
return _format_response(response, format=format, template='show_job.html')
@app.route('/jobs/<int:id>/results.<format>', methods=['GET'])
def get_results(id, format='json'):
"""
GET /jobs/<id>/results.<format>
Get job results by id.
Returns::
{
id: <job id>
status: 'PENDING|ACTIVE|COMPLETE|ERROR|UNKNOWN',
result: <job value> (absent if status != COMPLETE)
trace: <error trace> (absent if status != ERROR)
}
"""
response = SCHEDULER.results(id)
response['id'] = id
#print "returning response",response
return _format_response(response, format=format)
@app.route('/jobs/<int:id>/status.<format>', methods=['GET'])
def get_status(id, format='json'):
"""
GET /jobs/<id>/status.<format>
Get job status by id.
Returns::
{
id: <job id>,
status: 'PENDING|ACTIVE|COMPLETE|ERROR|UNKNOWN'
}
"""
response = { 'status': SCHEDULER.status(id) }
response['id'] = id
return _format_response(response, format=format)
@app.route('/jobs/<int:id>.<format>', methods=['DELETE'])
def delete_job(id, format='json'):
"""
DELETE /jobs/<id>.<format>
Deletes a job, returning the list of remaining jobs as <format>
"""
SCHEDULER.delete(id)
flash('Job %s deleted' % id)
response = dict(jobs=SCHEDULER.jobs())
return _format_response(response, format=format, template="list_jobs.html")
#return redirect(url_for('list_jobs', id=id, format=format))
@app.route('/jobs/nextjob.<format>', methods=['POST'])
def fetch_work(format='json'):
# TODO: verify signature
request = flask.request.json #@UndefinedVariable in request proxy
if request is None: flask.abort(415) # Unsupported media
job = SCHEDULER.nextjob(queue=request['queue'])
return _format_response(job, format=format)
@app.route('/jobs/<int:id>/postjob', methods=['POST'])
def return_work(id):
# TODO: verify signature corresponds to flask.request.form['queue']
# TODO: verify that work not already returned by another client
try:
#print "decoding <%s>"%flask.request.form['results']
results = json.loads(flask.request.form['results']) #@UndefinedVariable in request proxy
except:
import traceback;
logging.error(traceback.format_exc())
results = {
'status': 'ERROR',
'error': 'No results returned from the server',
'trace': flask.request.form['results'], #@UndefinedVariable in request proxy
}
_transfer_files(id)
SCHEDULER.postjob(id, results)
# Should be signaling code 204: No content
return _format_response({},format="json")
@app.route('/jobs/<int:id>/data/index.<format>')
def listfiles(id, format):
try:
path = store.path(id)
files = sorted(os.listdir(path))
finfo = [(f,os.path.getsize(os.path.join(path,f)))
for f in files if os.path.isfile(os.path.join(path,f))]
except:
finfo = []
response = { 'files': finfo }
response['id'] = id
return _format_response(response, format=format, template="index.html")
# TODO: don't allow putfiles without authentication
#@app.route('/jobs/<int:id>/data/', methods=['GET','PUT'])
def putfiles(id):
if flask.request.method=='PUT': #@UndefinedVariable in request proxy
# TODO: verify signature
_transfer_files(id)
return redirect(url_for('getfile',id=id,filename='index.html'))
@app.route('/jobs/<int:id>/data/<filename>')
def getfile(id, filename):
as_attachment = filename.endswith('.htm') or filename.endswith('.html')
if filename.endswith('.json'):
mimetype = "application/json"
else:
mimetype = None
return send_from_directory(store.path(id), filename,
mimetype=mimetype, as_attachment=as_attachment)
#@app.route('/jobs/<int:id>.<format>', methods=['PUT'])
#def update_job(id, format='.json'):
# """
# PUT /job/<id>.<format>
#
# Updates a job using data from the job submission form.
# """
# book = Book(id=id, name=u"I don't know") # Your query
# book.name = request.form['name'] # Save it
# flash('Book %s updated!' % book.name)
# return redirect(url_for('show_job', id=id, format=format))
#@app.route('/jobs/new.html')
#def new_job_form():
# """
# GET /jobs/new
#
# Returns a job submission form.
# """
# return render_template('new_job.html')
#@app.route('/jobss/<int:id>/edit.html')
#def edit_job_form(id):
# """
# GET /books/<id>/edit
#
# Form for editing job details
# """
# book = Book(id=id, name=u'Something crazy') # Your query
# return render_template('edit_book.html', book=book)
def _transfer_files(jobid):
logging.warn("XSS attacks possible if stored file is mimetype html")
for file in flask.request.files.getlist('file'): #@UndefinedVariable in request proxy
if not file: continue
filename = secure_filename(os.path.split(file.filename)[1])
# Because we used named temporary files that aren't deleted on
# closing as our streaming file type, we can simply move the
# resulting files to the store rather than copying them.
file.stream.close()
logging.warn("moving %s -> %s"%(file.stream.name, os.path.join(store.path(jobid),filename)))
os.rename(file.stream.name, os.path.join(store.path(jobid),filename))
def init_scheduler(conf):
if conf == 'slurm':
from . import slurm
Scheduler = slurm.Scheduler
elif conf == 'direct':
logging.warn("direct scheduler is not a good choice!")
try: os.nice(19)
except: pass
from . import simplequeue
Scheduler = simplequeue.Scheduler
elif conf == 'dispatch':
from . import dispatcher
Scheduler = dispatcher.Scheduler
else:
raise ValueError("unknown scheduler %s"%conf)
return Scheduler()
def serve():
app.run(host='0.0.0.0')
def fullpath(p): return os.path.abspath(os.path.expanduser(p))
def configure(jobstore=None, jobkey=None, jobdb=None, scheduler=None):
global SCHEDULER, app
if jobstore:
store.ROOT = fullpath(jobstore)
if jobkey:
app.config['SECRET_KEY'] = open(fullpath(jobkey)).read()
if jobdb:
from . import db
db.DB_URI = jobdb
SCHEDULER = init_scheduler(scheduler)
if __name__ == '__main__':
configure(jobstore='/tmp/server/%s',
jobdb='sqlite:///tmp/jobqueue.db',
jobkey='~/.bumps/key',
scheduler='dispatch',
)
app.config['DEBUG'] = True
serve()
|