jsonapi-client 0.9.7

客户端会话

fromjsonapi_clientimportSession,Filter,ResourceTuples=Session('http://localhost:8080/')# To start session in async modes=Session('http://localhost:8080/',enable_async=True)# You can also pass extra arguments that are passed directly to requests or aiohttp methods,# such as authentication objects=Session('http://localhost:8080/',request_kwargs=dict(auth=HttpBasicAuth('user','password'))# You can also use Session as a context manager. Changes are committed in the end# and session is closed.withSession(...)ass:yourcode# Or with enable_async=TrueasyncwithSession(...,enable_async=True):yourcode# If you are not using context manager, you need to close session manuallys.close()# Fetching documentsdocuments=s.get('resource_type')# Or if you want only 1, thendocuments=s.get('resource_type','id_of_document')# AsyncIO the same but remember to await:documents=awaits.get('resource_type')

过滤和包括

# You need first to specify your filter instance.# - filtering with two criteria (and)filter=Filter(attribute='something',attribute2='something_else')# - filtering some-dict.some-attr == 'something'filter=Filter(some_dict__some_attr='something'))# Same thing goes for including.# - including two fieldsinclude=Inclusion('related_field','other_related_field')# Custom syntax for request parameters.# If you have different URL schema for filtering or other GET parameters,# you can implement your own Modifier class (derive it from Modifier and# reimplement appended_query).modifier=Modifier('filter[post]=1&filter[author]=2')# All above classes subclass Modifier and can be added to concatenate# parametersmodifier_sum=filter+include+modifier# Now fetch your documentfiltered=s.get('resource_type',modifier_sum)# AsyncIO with await# To access resources included in document:r1=document.resources[0]# first ResourceObject of document.r2=document.resource# if there is only 1 resource we can use this

分页

# Pagination links can be accessed via Document object.next_doc=document.links.next.fetch()# AsyncIOnext_doc=awaitdocument.links.next.fetch()# Iteration through results (uses pagination):forrins.iterate('resource_type'):print(r)# AsyncIO:asyncforrins.iterate('resource_type'):print(r)

资源属性和关系访问

# - attribute accessattr1=r1.some_attrnested_attr=r1.some_dict.some_attr#   Attributes can always also be accessed via __getitem__:nested_attr=r1['some-dict']['some-attr']# If there is namespace collision, you can also access attributes via .fields proxy# (both attributes and relationships)attr2=r1.fields.some_attr# - relationship access.#   * Sync, this gives directly ResourceObjectrel=r1.some_relationattr3=r1.some_relation.some_attr# Relationship attribute can be accessed directly#   * AsyncIO, this gives Relationship object instead because we anyway need to#     call asynchronous fetch function.rel=r1.some_relation#     To access ResourceObject you need to first fetch contentawaitr1.some_relation.fetch()#     and then you can access associated resourceobjectres=r1.some_relation.resourceattr3=res.some_attr# Attribute access through ResourceObject# If you need to access relatinoship object itself (with sync API), you can do it via# .relationships proxy. For example, if you are interested in links or metadata# provided within relationship, or intend to manipulate relationship.rel_obj=r1.relationships.relation_name

资源更新

# Updating / patching existing resourcesr1.some_attr='something else'# Patching element in nested jsonr1.some_dict.some_dict.some_attr='something else'# change relationships, to-many. Accepts also iterable of ResourceObjects/# ResourceIdentifiers/ResourceTuplesr1.comments=['1','2']# or if resource type is not known or can have multiple types of resourcesr1.comments_or_people=[ResourceTuple('1','comments'),ResourceTuple('2','people')]# or if you want to add some resources you canr1.comments_or_people+=[ResourceTuple('1','people')]r1.commit()# change to-one relationshipsr1.author='3'# accepts also ResourceObjects/ResourceIdentifiers/ResourceTuple# or resource type is not known (via schema etc.)r1.author=ResourceTuple('3','people')# Committing changes (PATCH request)r1.commit(meta={'some_meta':'data'})# Resource committing supports optional meta data# AsyncIOawaitr1.commit(meta={'some_meta':'data'})

创建新资源

# Creating new resources. Schema must be given. Accepts dictionary of schema models# (key is model name and value is schema as json-schema.org).models_as_jsonschema={'articles':{'properties':{'title':{'type':'string'},'author':{'relation':'to-one','resource':['people']},'comments':{'relation':'to-many','resource':['comments']},}},'people':{'properties':{'first-name':{'type':'string'},'last-name':{'type':'string'},'twitter':{'type':['null','string']},}},'comments':{'properties':{'body':{'type':'string'},'author':{'relation':'to-one','resource':['people']}}}}# If you type schema by hand, it could be more convenient to type it as yml in a file# insteads=Session('http://localhost:8080/',schema=models_as_jsonschema)a=s.create('articles')# Creates empty ResourceObject of 'articles' typea.title='Test title'# Validates and performs POST request, and finally updates resource based on server responsea.commit(meta={'some_meta':'data'})# Or with AsyncIO, remember to awaitawaita.commit(meta={'some_meta':'data'})# Commit metadata could be also saved in advance:a.commit_metadata={'some_meta':'data'}# You can also commit all changed resources in session bys.commit()# or with AsyncIOawaits.commit()# Another example of resource creation, setting attributes and relationships & committing:# If you have underscores in your field names, you can pass them in fields keyword argument as# a dictionary:cust1=s.create_and_commit('articles',attribute='1',dict_object__attribute='2',to_one_relationship='3',to_many_relationship=['1','2'],fields={'some_field_with_underscore':'1'})# Async:cust1=awaits.create_and_commit('articles',attribute='1',dict_object__attribute='2',to_one_relationship='3',to_many_relationship=['1','2'],fields={'some_field_with_underscore':'1'})

删除资源

# Delete resourcecust1.delete()# Mark to be deletedcust1.commit()# Actually delete

0.9.7（2019-02-01）

• 在meta中支持getitem
• 处理空关系数据列表
• 允许以迭代器的形式返回到关系的链接
• 修复处理空的一对一关系
• 不要显式引用筛选器值
• 包括支持

0.9.6（2017-06-26）

• When creating new resources, use default value specified in jsonschema, when available.

0.9.5（2017-06-16）

• Change Session.create_and_commit signature similarly as Session.create

0.9.4（2017-06-16）

• Remove ? from filenames (illegal in Windows)
• Pass event loop aiohttp’s ClientSession
• Return resource from .commit if return status is 202
• Support underscores in field names in Session.create() through fields keyword argument.
• Add support for extra arguments such as authentication object
• AsyncIO support for context manager usage of Session

0.9.3（2017-04-03）

• Added aiohttp to install requirements

0.9.2（2017-04-03）

• Github release.

0.9.1（2017-03-23）

• Fix async content_type checking
• Use Python 3’s new typing.NamedTuple instead of collections.NamedTuple
• Make included resources available from Document
• ResourceObject.json property