mongodb.py
from chatterbot.storage import StorageAdapter class Query(object): def __init__(self, query={}):
self.query = query def value(self):
return self.query.copy() def raw(self, data):
query = self.query.copy() query.update(data) return Query(query) def statement_text_equals(self, statement_text):
query = self.query.copy() query['text'] = statement_text return Query(query) def statement_text_not_in(self, statements):
query = self.query.copy() if 'text' not in query:
query['text'] = {} if '$nin' not in query['text']:
query['text']['$nin'] = [] query['text']['$nin'].extend(statements) return Query(query) def statement_response_list_contains(self, statement_text):
query = self.query.copy() if 'in_response_to' not in query:
query['in_response_to'] = {} if '$elemMatch' not in query['in_response_to']:
query['in_response_to']['$elemMatch'] = {} query['in_response_to']['$elemMatch']['text'] = statement_text return Query(query) def statement_response_list_equals(self, response_list):
query = self.query.copy() query['in_response_to'] = response_list return Query(query) class MongoDatabaseAdapter(StorageAdapter):
"""
The MongoDatabaseAdapter is an interface that allows
ChatterBot to store statements in a MongoDB database. :keyword database: The name of the database you wish to connect to.
:type database: str .. code-block:: python database='chatterbot-database' :keyword database_uri: The URI of a remote instance of MongoDB.
:type database_uri: str .. code-block:: python database_uri='mongodb://example.com:8100/'
""" def __init__(self, **kwargs):
super(MongoDatabaseAdapter, self).__init__(**kwargs)
from pymongo import MongoClient
from pymongo.errors import OperationFailure self.database_name = self.kwargs.get(
'database', 'chatterbot-database'
)
self.database_uri = self.kwargs.get(
'database_uri', 'mongodb://localhost:27017/'
) # Use the default host and port
self.client = MongoClient(self.database_uri) # Increase the sort buffer to 42M if possible
try:
self.client.admin.command({'setParameter': 1, 'internalQueryExecMaxBlockingSortBytes': 44040192})
except OperationFailure:
pass # Specify the name of the database
self.database = self.client[self.database_name] # The mongo collection of statement documents
self.statements = self.database['statements'] # The mongo collection of conversation documents
self.conversations = self.database['conversations'] # Set a requirement for the text attribute to be unique
self.statements.create_index('text', unique=True) self.base_query = Query() def get_statement_model(self):
"""
Return the class for the statement model.
"""
from chatterbot.conversation import Statement # Create a storage-aware statement
statement = Statement
statement.storage = self return statement def get_response_model(self):
"""
Return the class for the response model.
"""
from chatterbot.conversation import Response # Create a storage-aware response
response = Response
response.storage = self return response def count(self):
return self.statements.count() def find(self, statement_text):
Statement = self.get_model('statement')
query = self.base_query.statement_text_equals(statement_text) values = self.statements.find_one(query.value()) if not values:
return None del values['text'] # Build the objects for the response list
values['in_response_to'] = self.deserialize_responses(
values.get('in_response_to', [])
) return Statement(statement_text, **values) def deserialize_responses(self, response_list):
"""
Takes the list of response items and returns
the list converted to Response objects.
"""
Statement = self.get_model('statement')
Response = self.get_model('response')
proxy_statement = Statement('') for response in response_list:
text = response['text']
del response['text'] proxy_statement.add_response(
Response(text, **response)
) return proxy_statement.in_response_to def mongo_to_object(self, statement_data):
"""
Return Statement object when given data
returned from Mongo DB.
"""
Statement = self.get_model('statement')
statement_text = statement_data['text']
del statement_data['text'] statement_data['in_response_to'] = self.deserialize_responses(
statement_data.get('in_response_to', [])
) return Statement(statement_text, **statement_data) def filter(self, **kwargs):
"""
Returns a list of statements in the database
that match the parameters specified.
"""
import pymongo query = self.base_query order_by = kwargs.pop('order_by', None) # Convert Response objects to data
if 'in_response_to' in kwargs:
serialized_responses = []
for response in kwargs['in_response_to']:
serialized_responses.append({'text': response}) query = query.statement_response_list_equals(serialized_responses)
del kwargs['in_response_to'] if 'in_response_to__contains' in kwargs:
query = query.statement_response_list_contains(
kwargs['in_response_to__contains']
)
del kwargs['in_response_to__contains'] query = query.raw(kwargs) matches = self.statements.find(query.value()) if order_by: direction = pymongo.ASCENDING # Sort so that newer datetimes appear first
if order_by == 'created_at':
direction = pymongo.DESCENDING matches = matches.sort(order_by, direction) results = [] for match in list(matches):
results.append(self.mongo_to_object(match)) return results def update(self, statement):
from pymongo import UpdateOne
from pymongo.errors import BulkWriteError data = statement.serialize() operations = [] update_operation = UpdateOne(
{'text': statement.text},
{'$set': data},
upsert=True
)
operations.append(update_operation) # Make sure that an entry for each response is saved
for response_dict in data.get('in_response_to', []):
response_text = response_dict.get('text') # $setOnInsert does nothing if the document is not created
update_operation = UpdateOne(
{'text': response_text},
{'$set': response_dict},
upsert=True
)
operations.append(update_operation) try:
self.statements.bulk_write(operations, ordered=False)
except BulkWriteError as bwe:
# Log the details of a bulk write error
self.logger.error(str(bwe.details)) return statement def create_conversation(self):
"""
Create a new conversation.
"""
conversation_id = self.conversations.insert_one({}).inserted_id
return conversation_id def get_latest_response(self, conversation_id):
"""
Returns the latest response in a conversation if it exists.
Returns None if a matching conversation cannot be found.
"""
from pymongo import DESCENDING statements = list(self.statements.find({
'conversations.id': conversation_id
}).sort('conversations.created_at', DESCENDING)) if not statements:
return None return self.mongo_to_object(statements[-2]) def add_to_conversation(self, conversation_id, statement, response):
"""
Add the statement and response to the conversation.
"""
from datetime import datetime, timedelta
self.statements.update_one(
{
'text': statement.text
},
{
'$push': {
'conversations': {
'id': conversation_id,
'created_at': datetime.utcnow()
}
}
}
)
self.statements.update_one(
{
'text': response.text
},
{
'$push': {
'conversations': {
'id': conversation_id,
# Force the response to be at least one millisecond after the input statement
'created_at': datetime.utcnow() + timedelta(milliseconds=1)
}
}
}
) def get_random(self):
"""
Returns a random statement from the database
"""
from random import randint count = self.count() if count < 1:
raise self.EmptyDatabaseException() random_integer = randint(0, count - 1) statements = self.statements.find().limit(1).skip(random_integer) return self.mongo_to_object(list(statements)[0]) def remove(self, statement_text):
"""
Removes the statement that matches the input text.
Removes any responses from statements if the response text matches the
input text.
"""
for statement in self.filter(in_response_to__contains=statement_text):
statement.remove_response(statement_text)
self.update(statement) self.statements.delete_one({'text': statement_text}) def get_response_statements(self):
"""
Return only statements that are in response to another statement.
A statement must exist which lists the closest matching statement in the
in_response_to field. Otherwise, the logic adapter may find a closest
matching statement that does not have a known response.
"""
response_query = self.statements.aggregate([{'$group': {'_id': '$in_response_to.text'}}]) responses = []
for r in response_query:
try:
responses.extend(r['_id'])
except TypeError:
pass _statement_query = {
'text': {
'$in': responses
}
} _statement_query.update(self.base_query.value())
statement_query = self.statements.find(_statement_query)
statement_objects = []
for statement in list(statement_query):
statement_objects.append(self.mongo_to_object(statement))
return statement_objects def drop(self):
"""
Remove the database.
"""
self.client.drop_database(self.database_name)
mongodb.py的更多相关文章
- Python mongoDB 的简单操作
#!/usr/bin/env python # coding:utf-8 # Filename:mongodb.py from pymongo import MongoClient,ASCENDING ...
- Python连接MongoDB数据库并执行操作
原文:https://blog.51cto.com/1767340368/2092813 环境设置: [root@mongodb ~]# cat /etc/redhat-release CentOS ...
- MongoDB与python 交互
一.安装pymongo 注意 :当同时安装了python2和python3,为区分两者的pip,分别取名为pip2和pip3. 推荐:https://www.cnblogs.com/thunderLL ...
- 2019-05-25 Python之Mongodb的使用
Mongodb学习总结: one.插入数据pass two.查看数据pass three.修改数据pass four.数据排序pass five.删除数据pass 一.安装Mongodb https: ...
- ansible的lookup
lookup路径: /usr/lib/python2.7/site-packages/ansible/plugins/lookup 所有的lookup插件列表cartesian.py dnstxt.p ...
- SaltStack介绍及简单配置-第一篇
SaltStack介绍 一种全新的基础设施管理方式,部署轻松,在几分钟内可运行起来,扩展性好,很容易管理上万台服务器,速度够快,服务器之间秒级通讯. salt底层采用动态的连接总线, 使其可以用于编配 ...
- appium 爬取抖音
1.MongoDB.py import pymongo from pymongo.collection import Collection client = pymongo.MongoClient(h ...
- Scrapy的piplines.py存储文件和存储mongodb
一.将数据保存到文件 1.piplines.py文件 import json class TencentPipeline(object): def open_spider(self,spider): ...
- py操作mongodb总结
python使用的版本 python3. python操作mongodb使用的是pymongo,安装方法: pip install pymongo 测试 PyMongo 接下来我们可以创建一个测试文件 ...
随机推荐
- C++ primer分章节快速回顾
第三章: 1,sozeof(int): int n_int=INT_MAX; sizeof n_int;(对变量括号可选) 2,#include<climits>包含一些类型的最大值3,c ...
- BZOJ——3412: [Usaco2009 Dec]Music Notes乐谱
http://www.lydsy.com/JudgeOnline/problem.php?id=3412 Time Limit: 3 Sec Memory Limit: 128 MBSubmit: ...
- POJ 3373 Changing Digits 记忆化搜索
这道题我是看了别人的题解才做出来的.题意和题解分析见原文http://blog.csdn.net/lyy289065406/article/details/6698787 这里写一下自己对题目的理解. ...
- bzoj 5123: [Lydsy1712月赛]线段树的匹配
设f[0/1][x]为区间[1,x]的根向下 不选(0)或者选(1) 的dp pair<最优值,方案数>. 可以很容易的发现总状态数就是log级别的,因为2*n 与 (2*n+1 或者 ...
- Linux基础学习3
鸟哥私房菜第十一章 read 用来读取键盘输入的变量,常被用在shell script的交互当中 [root@www ~]# read [-pt] variable 选项不参数: -p :后面可以 ...
- 关于Web应用和容器的指纹收集以及自动化软件的制作
一次对Web应用的渗透,九成都是从信息收集开始,所以信息收集就显得尤为重要.关键信息的收集可以使你在后期渗透的时候更加的得心应手,把渗透比喻成走黑暗迷宫的话,那信息收集可以帮你点亮迷宫的大部分地图. ...
- RTC实时时钟驱动
RTC(Real-Time Clock)实时时钟为操作系统提供了一个可靠的时间,并且在断电的情况下,RTC实时时钟也可以通过电池供电,一直运行下去. RTC通过STRB/LDRB这两个ARM指令向CP ...
- java基础篇2之枚举
1.为什么要有枚举 问题:要定义星期几或者性别的变量,该怎么定义? 假设用1-7分别表示星期一到星期日,但有人可能会写成int weekday=0; 枚举就是要让某个类型的变量的取值只能为若干个固定值 ...
- 【小程序】微信小程序开发实践
帐号相关流程 注册范围 企业 政府 媒体 其他组织 换句话讲就是不让个人开发者注册. :) 填写企业信息 不能使用和之前的公众号账户相同的邮箱,也就是说小程序是和微信公众号一个层级的. 填写公司机构信 ...
- koajs 项目实战(二)
此篇文章,接 koajs 项目实战(一)后写 (六)表单提交和参数接收 表单: <form method="post" action="./users/zhuce& ...