def load_extract(body, ctype):
'''
Read two JSON files merged with a LINEFEED, and use the specification
to extract an array from the JSON source
Sample queries:
* curl --request POST --data-binary @- http://localhost:8880/load.extract.json < test/data/load/json_plus_spec.dat
'''
file1, file2 = body.split('\f')
file1_parsed = []
if ejsonify.is_json(file1, file1_parsed):
file1_parsed = file1_parsed[0]
file2_parsed = []
if not file2 or ejsonify.is_json(file2, file2_parsed):
if file2:
file2_parsed = file2_parsed[0]
else:
file2_parsed = None
result = ejsonify.pull_ejson_by_patterns(file1_parsed, file2_parsed)
items = result[u'items']
profile = add_ejson_profile(items)
result = json.dumps({'items': items, 'data_profile': profile}, indent=4)
return result
raise ValueError('Unable to process content')
def json_nav_prep(body, ctype):
'''
Read and analyze a JSON file to support the UI for having the user
navigate and select an array for extract (see load_extract)
Sample queries:
* curl --request POST --data-binary @- http://localhost:8880/json.nav.prep.json < test/data/load/somefeed.json
'''
ejsonify_output = []
if ejsonify.is_json(body, ejsonify_output):
obj = ejsonify_output[0]
result = ejsonify.analyze_for_nav(obj)
result = json.dumps(result, indent=4)
return result
else:
raise ValueError('Unable to process content')
else:
lead = body.strip()
if lead.startswith('%%') or lead.startswith('@'):
#BibTex
cmdline = 'bib2xml'
process = Popen(cmdline, stdin=PIPE, stdout=PIPE, universal_newlines=True, shell=True)
mods, perr = process.communicate(input=body)
if not mods:
#FIXME: L10N
raise ValueError('Empty output from the command line. Probably a failure. Command line: "%s"'%cmdline)
#print >> sys.stderr, mods[:100]
data, diag_info = mods2json(body, diagnostics)
imt = 'application/x-bibtex'
else:
ejsonify_output = []
if ejsonify.is_json(body, ejsonify_output):
obj = ejsonify_output[0]
data = obj[u'items']
fixup_obj_labels = False
imt = 'application/json'
else:
#FIXME: how to deal with CSV character sets?
data = readcsv(body)
if maxcount:
data = data[:int(maxcount)]
#FIXME: Following redundant now that we have add_ejson_profile
#Keeping it to avoid breaking what's working for now
objkeys = dict([ (k, k) for obj in data for k in obj ])
#FIXME: reduce from 3 full passes through obj to 2 (don't think we can go lower than 2)
