replace duanwei with 儿童

replace_content_in_mongo.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
import logging
import pymongo


NEW_HASH = '0f533b43e264260e17fed6a15e05ffd6c3f8e3b1'
HASH_TEXT = 'duanwei-儿童'
OLD_HASH_LS = ['efacd9afc28bed0712c4e904585f10a277fbfc7d',
               '59f420d90a824716eb07df9b698aa2ecedf3f0dc',
               '558cddba6639b9de3c5dbb0cc5953b0c485ca349',
               '51ba3c692aca1c42b5fbf9860325450c226a13ea']


log = logging.getLogger(__name__)


def run():
    conn = pymongo.Connection(host='192.168.1.202', port=27017, safe=True)
    # conn = pymongo.Connection(safe=True)
    # db = conn['tss-clean-data-0920']
    db = conn['tss-dev']
    subjects = db['subjects']
    layers = db['layers']
    metadata = db['metadata']

    # metadata
    log.info('metadata')
    meta_docus = metadata.find(
        {
            'subjectFeatures.' + OLD_HASH_LS[0]: {'$exists': 1}
        }
    )

    for docu in meta_docus:
        for feature in OLD_HASH_LS:
            if feature in docu['subjectFeatures']:
                del docu['subjectFeatures'][feature]

            if feature in docu['booleanSubjectFeaturesGroups']['duanwei']:
                docu['booleanSubjectFeaturesGroups']['duanwei'].remove(feature)

        docu['subjectFeatures'][NEW_HASH] = HASH_TEXT

        if NEW_HASH not in docu['booleanSubjectFeaturesGroups']['duanwei']:
            docu['booleanSubjectFeaturesGroups']['duanwei'].append(NEW_HASH)

        del docu['_id']

        metadata.update(
            {'name': docu['name']},
            {'$set': docu},
            multi=True
        )

    # subjects
    log.info('subjects')
    sbj_docus = subjects.find(
        {
            'subjectFeatures.' + OLD_HASH_LS[0]: {'$exists': 1}
        }
    )

    for docu in sbj_docus:
        sum_val = 0
        for feature in OLD_HASH_LS:
            if feature in docu['subjectFeatures']:
                sum_val += docu['subjectFeatures'][feature]
                del docu['subjectFeatures'][feature]

        del docu['_id']

        docu['subjectFeatures'][NEW_HASH] = sum_val
        subjects.update(
            {'set': docu['set'], 'layer': docu['layer'], 'sid': docu['sid']},
            {'$set': docu},
            multi=True
        )

    # layers
    log.info('layers')
    layer_docus = layers.find(
        {'subjectFeatures': {'$in': OLD_HASH_LS}}
    )

    for docu in layer_docus:
        for feature in OLD_HASH_LS:
            if feature in docu['subjectFeatures']:
                docu['subjectFeatures'].remove(feature)

        if NEW_HASH not in docu['subjectFeatures']:
            docu['subjectFeatures'].append(NEW_HASH)

        del docu['_id']

        layers.update(
            {'set': docu['set'], 'layer': docu['layer']},
            {'$set': docu},
            multi=True
        )


if __name__ == '__main__':
    run()