有没有办法删除太多的其他条件？

Question

我目前正在制作一个使用python的交互式系统，它能够理解和回复。因此，对于机器分析和处理有很多条件。例如。采取下面的代码（仅供参考）：

if ('goodbye') in message:       
     rand = ['Goodbye Sir', 'Jarvis powering off in 3, 2, 1, 0'] 
     speekmodule.speek(rand,n,mixer) 
     break 

    if ('hello') in message or ('hi') in message: 
     rand = ['Wellcome to Jarvis virtual intelligence project. At your service sir.'] 
     speekmodule.speek(rand,n,mixer) 

    if ('thanks') in message or ('tanks') in message or ('thank you') in message: 
     rand = ['You are wellcome', 'no problem'] 
     speekmodule.speek(rand,n,mixer) 

    if message == ('jarvis'): 
     rand = ['Yes Sir?', 'What can I doo for you sir?'] 
     speekmodule.speek(rand,n,mixer) 

    if ('how are you') in message or ('and you') in message or ('are you okay') in message: 
     rand = ['Fine thank you'] 
     speekmodule.speek(rand,n,mixer) 

    if ('*') in message: 
     rand = ['Be polite please'] 
     speekmodule.speek(rand,n,mixer) 

    if ('your name') in message: 
     rand = ['My name is Jarvis, at your service sir'] 
     speekmodule.speek(rand,n,mixer) 

那么，有没有一种方式，我可以取代所有这些，如果其他条件？因为条件会更多，并且会使执行速度变慢。

Answer 1

做一个独特的 “如果”：

if 'goodbye' in message:       
    rand = ['Goodbye Sir', 'Jarvis powering off in 3, 2, 1, 0'] 

elif 'hello' in message or 'hi' in message: 
    rand = ['Wellcome to Jarvis virtual intelligence project. At your service sir.'] 

elif 'thanks' in message or 'tanks' in message or ('thank you') in message: 
    rand = ['You are wellcome', 'no problem'] 

elif message == 'jarvis': 
    rand = ['Yes Sir?', 'What can I doo for you sir?'] 

elif 'how are you' in message or 'and you' in message or ('are you okay') in message: 
    rand = ['Fine thank you'] 

elif '*' in message: 
    rand = ['Be polite please'] 

elif 'your name' in message: 
    rand = ['My name is Jarvis, at your service sir'] 

else: 
    raise NotImplementedError("What to do?") 

speekmodule.speek(rand, n, mixer) 

用正则表达式的映射：

mapping = { 
    r"\bgoodbye\b": ['Goodbye Sir', 'Jarvis powering off in 3, 2, 1, 0'], 
    r"\bhello\b": ['Wellcome to Jarvis virtual intelligence project. At your service sir.'], 
    ...} 

for regex, rand in mapping.items(): 
    if re.search(message, flags=re.I): 
     break 
else: 
    raise NotImplementedError("What to do?") 
speekmodule.speek(rand, n, mixer) 

它是由你来决定。

Answer 2

if/elif/else是一种在Python中构造这种代码的自然方式。正如@imant指出的那样，在简单分支的情况下，您可以使用基于字典的方法，但我在您的谓词中看到一些复杂的逻辑，因此您必须在任何情况下检查所有谓词，并且不会有任何性能获得另一个代码结构。

虽然它可能是一个有点更具可读性和更容易，如果你分解出你的谓词和操作这样的维护：

from collections import OrderedDict 

def goodbye_p(message): 
    return 'goodbye' in message 

def goodbye_a(): 
    rand = ['Goodbye Sir', 'Jarvis powering off in 3, 2, 1, 0'] 
    # As @Moinuddin Quadri I also assume that your `speek` method 
    # says random message from a list. 
    # Otherwise you can use `random.choice` method 
    # to get a random message out of a list: `random.choice(messages)`. 
    speekmodule.speek(rand, n, mixer) 


def hello_p(message): 
    return 'hello' in message or 'hi' in message 

def hello_a(): 
    rand = ['Wellcome to Jarvis virtual intelligence project. At your service sir.'] 
    speekmodule.speek(rand, n, mixer) 

# Use `OrderedDict` instead of `dict` to control order 
# of checks and actions. 
branches = OrderedDict([ 
    # (predicate as key, action as value) 
    (goodbye_p, goodbye_a), 
    (hello_p, hello_a), 
]) 

for predicate, action in branches.items(): 
    if predicate(message): 
     action_result = action() 
     # You can add some logic here based on action results. 
     # E.g. you can return some special object from `goodbye_a` 
     # and then shut down Jarvis here. 
     # Or if your actions are exclusive, you can add `break` here. 

如果所有的谓词是相同的，并且只包含子串检查，然后将其可能会更高效地将元组（例如('hello', 'hi')）作为字典键。然后你就可以在这些键这样的迭代：

for words, action in branches.items(): 
    if any(word in message for word in words): 
     action() 

Answer 3

首先创建一个键一个dict对象作为字符串的tuple要匹配你的message，它与你的贾维斯是假设响应值string关联。例如：

jarvis_dict = { 
    ('goodbye',) : ['Goodbye Sir', 'Jarvis powering off in 3, 2, 1, 0'], 
    ('hello', 
    'hi')  : ['Wellcome to Jarvis virtual intelligence project. At your service sir.'], 
    ('thanks', 
    'tanks', 
    'thank you') : ['You are wellcome', 'no problem'], 
    ('jarvis',) : ['Yes Sir?', 'What can I doo for you sir?'], 
    ('how are you', 
    'and you', 
    'are you okay'): ['Fine thank you'], 
    ('*',)  : ['Be polite please'], 
    ('your name',): ['My name is Jarvis, at your service sir'] 
} 

现在迭代的你dict每个键来检查任何子串是否是消息的一部分，如果有任何匹配，调用speekmodule.speek(rand,n,mixer)功能：

for key, value in jarvis_dict.items(): 
    if any(item in message for item in key): 
     speekmodule.speek(value, n, mixer) 

注意：这里我假设您的代码中的speekmodule.speek(value, n, mixer)正在工作，因为您的代码中没有关于该声明的信息。我刚刚将rand替换为value，因为它与您的代码中使用的dict返回的str列表相同。

Answer 4

使用字典：

someCollections = { 
    'goodbye': "Something1", 
    'hello': "Somthing2", 
    ... 
} 

speekmodule(someCollections [SomeKey],...)