3. Basic CouchDB API: couchdb.client

Python client API for CouchDB.

>>> server = Server()
>>> db = server.create('python-tests')
>>> doc_id, doc_rev = db.save({'type': 'Person', 'name': 'John Doe'})
>>> doc = db[doc_id]
>>> doc['type']
u'Person'
>>> doc['name']
u'John Doe'
>>> del db[doc.id]
>>> doc.id in db
False

>>> del server['python-tests']

3.1. Server

class couchdb.client.Server(url='http://localhost:5984/', full_commit=True, session=None)

Representation of a CouchDB server.

>>> server = Server() # connects to the local_server
>>> remote_server = Server('http://example.com:5984/')
>>> secure_remote_server = Server('https://username:password@example.com:5984/')

This class behaves like a dictionary of databases. For example, to get a list of database names on the server, you can simply iterate over the server object.

New databases can be created using the create method:

>>> db = server.create('python-tests')
>>> db
<Database 'python-tests'>

You can access existing databases using item access, specifying the database name as the key:

>>> db = server['python-tests']
>>> db.name
'python-tests'

Databases can be deleted using a del statement:

>>> del server['python-tests']

add_user(name, password, roles=None)

Add regular user in authentication database.

Parameters:

• name – name of regular user, normally user id
• password – password of regular user
• roles – roles of regular user

Returns:

(id, rev) tuple of the registered user

Return type:

tuple

config()

The configuration of the CouchDB server.

The configuration is represented as a nested dictionary of sections and options from the configuration files of the server, or the default values for options that are not explicitly configured.

Return type:dict

create(name)

Create a new database with the given name.

Parameters:name – the name of the database Returns:a Database object representing the created database Return type:Database Raises:PreconditionFailed – if a database with that name already exists

delete(name)

Delete the database with the specified name.

Parameters:name – the name of the database Raises:ResourceNotFound – if a database with that name does not exist Since:0.6

login(name, password)

Login regular user in couch db

Parameters:

• name – name of regular user, normally user id
• password – password of regular user

Returns:

authentication token

logout(token)

Logout regular user in couch db

Parameters:token – token of login user Returns:True if successfully logout Return type:bool

remove_user(name)

Remove regular user in authentication database.

Parameters:name – name of regular user, normally user id

replicate(source, target, **options)

Replicate changes from the source database to the target database.

Parameters:

• source – URL of the source database
• target – URL of the target database
• options – optional replication args, e.g. continuous=True

stats(name=None)

Server statistics.

Parameters:name – name of single statistic, e.g. httpd/requests (None – return all statistics)

tasks()

A list of tasks currently active on the server.

uuids(count=None)

Retrieve a batch of uuids

Parameters:count – a number of uuids to fetch (None – get as many as the server sends) Returns:a list of uuids

verify_token(token)

Verify user token

Parameters:token – authentication token Returns:True if authenticated ok Return type:bool

version()

The version string of the CouchDB server.

Note that this results in a request being made, and can also be used to check for the availability of the server.

Return type:unicode

version_info()

The version of the CouchDB server as a tuple of ints.

Note that this results in a request being made only at the first call. Afterwards the result will be cached.

Return type:tuple(int, int, int)

3.2. Database

class couchdb.client.Database(url, name=None, session=None)

Representation of a database on a CouchDB server.

>>> server = Server()
>>> db = server.create('python-tests')

New documents can be added to the database using the save() method:

>>> doc_id, doc_rev = db.save({'type': 'Person', 'name': 'John Doe'})

This class provides a dictionary-like interface to databases: documents are retrieved by their ID using item access

>>> doc = db[doc_id]
>>> doc                 
<Document u'...'@... {...}>

Documents are represented as instances of the Row class, which is basically just a normal dictionary with the additional attributes id and rev:

>>> doc.id, doc.rev     
(u'...', ...)
>>> doc['type']
u'Person'
>>> doc['name']
u'John Doe'

To update an existing document, you use item access, too:

>>> doc['name'] = 'Mary Jane'
>>> db[doc.id] = doc

The save() method creates a document with a random ID generated by CouchDB (which is not recommended). If you want to explicitly specify the ID, you’d use item access just as with updating:

>>> db['JohnDoe'] = {'type': 'person', 'name': 'John Doe'}

>>> 'JohnDoe' in db
True
>>> len(db)
2

>>> del server['python-tests']

If you need to connect to a database with an unverified or self-signed SSL certificate, you can re-initialize your ConnectionPool as follows (only applicable for Python 2.7.9+):

>>> db.resource.session.disable_ssl_verification()

changes(**opts)

Retrieve a changes feed from the database.

Parameters:opts – optional query string parameters Returns:an iterable over change notification dicts

cleanup()

Clean up old design document indexes.

Remove all unused index files from the database storage area.

Returns:a boolean to indicate successful cleanup initiation Return type:bool

commit()

If the server is configured to delay commits, or previous requests used the special X-Couch-Full-Commit: false header to disable immediate commits, this method can be used to ensure that any non-committed changes are committed to physical storage.

compact(ddoc=None)

Compact the database or a design document’s index.

Without an argument, this will try to prune all old revisions from the database. With an argument, it will compact the index cache for all views in the design document specified.

Returns:a boolean to indicate whether the compaction was initiated successfully Return type:bool

copy(src, dest)

Copy the given document to create a new document.

Parameters:

• src – the ID of the document to copy, or a dictionary or Document object representing the source document.
• dest – either the destination document ID as string, or a dictionary or Document instance of the document that should be overwritten.

Returns:

the new revision of the destination document

Return type:

str

Since:

0.6

create(data)

Create a new document in the database with a random ID that is generated by the server.

Note that it is generally better to avoid the create() method and instead generate document IDs on the client side. This is due to the fact that the underlying HTTP POST method is not idempotent, and an automatic retry due to a problem somewhere on the networking stack may cause multiple documents being created in the database.

To avoid such problems you can generate a UUID on the client side. Python (since version 2.5) comes with a uuid module that can be used for this:

from uuid import uuid4
doc_id = uuid4().hex
db[doc_id] = {'type': 'person', 'name': 'John Doe'}

Parameters:data – the data to store in the document Returns:the ID of the created document Return type:unicode

delete(doc)

Delete the given document from the database.

Use this method in preference over __del__ to ensure you’re deleting the revision that you had previously retrieved. In the case the document has been updated since it was retrieved, this method will raise a ResourceConflict exception.

>>> server = Server()
>>> db = server.create('python-tests')

>>> doc = dict(type='Person', name='John Doe')
>>> db['johndoe'] = doc
>>> doc2 = db['johndoe']
>>> doc2['age'] = 42
>>> db['johndoe'] = doc2
>>> db.delete(doc) 
Traceback (most recent call last):
  ...
ResourceConflict: (u'conflict', u'Document update conflict.')

>>> del server['python-tests']

Parameters:doc – a dictionary or Document object holding the document data Raises:ResourceConflict – if the document was updated in the database Since:0.4.1

delete_attachment(doc, filename)

Delete the specified attachment.

Note that the provided doc is required to have a _rev field. Thus, if the doc is based on a view row, the view row would need to include the _rev field.

Parameters:

• doc – the dictionary or Document object representing the document that the attachment belongs to
• filename – the name of the attachment file

Since:

0.4.1

explain(mango_query)

Explain a mango find-query.

Note: only available for CouchDB version >= 2.0.0

More information on the mango_query structure can be found here:

http://docs.couchdb.org/en/master/api/database/find.html#db-explain

>>> server = Server()
>>> db = server.create('python-tests')
>>> db['johndoe'] = dict(type='Person', name='John Doe')
>>> db['maryjane'] = dict(type='Person', name='Mary Jane')
>>> db['gotham'] = dict(type='City', name='Gotham City')
>>> mango = {'selector': {'type': 'Person'}, 'fields': ['name']}
>>> db.explain(mango)                          
{...}
>>> del server['python-tests']

Parameters:mango_query – a dict describing criteria used to select documents Returns:the query results as a list of Document (or whatever wrapper returns) Return type:dict

find(mango_query, wrapper=None)

Execute a mango find-query against the database.

Note: only available for CouchDB version >= 2.0.0

More information on the mango_query structure can be found here:

http://docs.couchdb.org/en/master/api/database/find.html#find-selectors

>>> server = Server()
>>> db = server.create('python-tests')
>>> db['johndoe'] = dict(type='Person', name='John Doe')
>>> db['maryjane'] = dict(type='Person', name='Mary Jane')
>>> db['gotham'] = dict(type='City', name='Gotham City')
>>> mango = {'selector': {'type': 'Person'},
...          'fields': ['name'],
...          'sort':[{'name': 'asc'}]}
>>> for row in db.find(mango):                          
...    print(row['name'])                               
John Doe
Mary Jane
>>> del server['python-tests']

Parameters:

• mango_query – a dictionary describing criteria used to select documents
• wrapper – an optional callable that should be used to wrap the resulting documents

Returns:

the query results as a list of Document (or whatever wrapper returns)

get(id, default=None, **options)

Return the document with the specified ID.

Parameters:

• id – the document ID
• default – the default value to return when the document is not found

Returns:

a Row object representing the requested document, or None if no document with the ID was found

Return type:

Document

get_attachment(id_or_doc, filename, default=None)

Return an attachment from the specified doc id and filename.

Parameters:

• id_or_doc – either a document ID or a dictionary or Document object representing the document that the attachment belongs to
• filename – the name of the attachment file
• default – default value to return when the document or attachment is not found

Returns:

a file-like object with read and close methods, or the value of the default argument if the attachment is not found

Since:

0.4.1

index()

Get an object to manage the database indexes.

Returns:an Indexes object to manage the databes indexes Return type:Indexes

info(ddoc=None)

Return information about the database or design document as a dictionary.

Without an argument, returns database information. With an argument, return information for the given design document.

The returned dictionary exactly corresponds to the JSON response to a GET request on the database or design document’s info URI.

Returns:a dictionary of database properties Return type:dict Since:0.4

iterview(name, batch, wrapper=None, **options)

Iterate the rows in a view, fetching rows in batches and yielding one row at a time.

Since the view’s rows are fetched in batches any rows emitted for documents added, changed or deleted between requests may be missed or repeated.

Parameters:

• name – the name of the view; for custom views, use the format design_docid/viewname, that is, the document ID of the design document and the name of the view, separated by a slash.
• batch – number of rows to fetch per HTTP request.
• wrapper – an optional callable that should be used to wrap the result rows
• options – optional query string parameters

Returns:

row generator

list(name, view, **options)

Format a view using a ‘list’ function.

Parameters:

• name – the name of the list function in the format designdoc/listname
• view – the name of the view in the format designdoc/viewname
• options – optional query string parameters

Returns:

(headers, body) tuple, where headers is a dict of headers returned from the list function and body is a readable file-like instance

name

The name of the database.

Note that this may require a request to the server unless the name has already been cached by the info() method.

Return type:basestring

purge(docs)

Perform purging (complete removing) of the given documents.

Uses a single HTTP request to purge all given documents. Purged documents do not leave any meta-data in the storage and are not replicated.

put_attachment(doc, content, filename=None, content_type=None)

Create or replace an attachment.

Note that the provided doc is required to have a _rev field. Thus, if the doc is based on a view row, the view row would need to include the _rev field.

Parameters:

• doc – the dictionary or Document object representing the document that the attachment should be added to
• content – the content to upload, either a file-like object or a string
• filename – the name of the attachment file; if omitted, this function tries to get the filename from the file-like object passed as the content argument value
• content_type – content type of the attachment; if omitted, the MIME type is guessed based on the file name extension

Since:

0.4.1

query(map_fun, reduce_fun=None, language='javascript', wrapper=None, **options)

Execute an ad-hoc query (a “temp view”) against the database.

Note: not supported for CouchDB version >= 2.0.0

>>> server = Server()
>>> db = server.create('python-tests')
>>> db['johndoe'] = dict(type='Person', name='John Doe')
>>> db['maryjane'] = dict(type='Person', name='Mary Jane')
>>> db['gotham'] = dict(type='City', name='Gotham City')
>>> map_fun = '''function(doc) {
...     if (doc.type == 'Person')
...         emit(doc.name, null);
... }'''
>>> for row in db.query(map_fun):
...     print(row.key)
John Doe
Mary Jane

>>> for row in db.query(map_fun, descending=True):
...     print(row.key)
Mary Jane
John Doe

>>> for row in db.query(map_fun, key='John Doe'):
...     print(row.key)
John Doe

>>> del server['python-tests']

Parameters:

• map_fun – the code of the map function
• reduce_fun – the code of the reduce function (optional)
• language – the language of the functions, to determine which view server to use
• wrapper – an optional callable that should be used to wrap the result rows
• options – optional query string parameters

Returns:

the view results

Return type:

ViewResults

revisions(id, **options)

Return all available revisions of the given document.

Parameters:id – the document ID Returns:an iterator over Document objects, each a different revision, in reverse chronological order, if any were found

save(doc, **options)

Create a new document or update an existing document.

If doc has no _id then the server will allocate a random ID and a new document will be created. Otherwise the doc’s _id will be used to identify the document to create or update. Trying to update an existing document with an incorrect _rev will raise a ResourceConflict exception.

Note that it is generally better to avoid saving documents with no _id and instead generate document IDs on the client side. This is due to the fact that the underlying HTTP POST method is not idempotent, and an automatic retry due to a problem somewhere on the networking stack may cause multiple documents being created in the database.

To avoid such problems you can generate a UUID on the client side. Python (since version 2.5) comes with a uuid module that can be used for this:

from uuid import uuid4
doc = {'_id': uuid4().hex, 'type': 'person', 'name': 'John Doe'}
db.save(doc)

Parameters:

• doc – the document to store
• options – optional args, e.g. batch=’ok’

Returns:

(id, rev) tuple of the save document

Return type:

tuple

show(name, docid=None, **options)

Call a ‘show’ function.

Parameters:

• name – the name of the show function in the format designdoc/showname
• docid – optional ID of a document to pass to the show function.
• options – optional query string parameters

Returns:

(headers, body) tuple, where headers is a dict of headers returned from the show function and body is a readable file-like instance

update(documents, **options)

Perform a bulk update or insertion of the given documents using a single HTTP request.

>>> server = Server()
>>> db = server.create('python-tests')
>>> for doc in db.update([
...     Document(type='Person', name='John Doe'),
...     Document(type='Person', name='Mary Jane'),
...     Document(type='City', name='Gotham City')
... ]):
...     print(repr(doc)) 
(True, u'...', u'...')
(True, u'...', u'...')
(True, u'...', u'...')

>>> del server['python-tests']

The return value of this method is a list containing a tuple for every element in the documents sequence. Each tuple is of the form (success, docid, rev_or_exc), where success is a boolean indicating whether the update succeeded, docid is the ID of the document, and rev_or_exc is either the new document revision, or an exception instance (e.g. ResourceConflict) if the update failed.

If an object in the documents list is not a dictionary, this method looks for an items() method that can be used to convert the object to a dictionary. Effectively this means you can also use this method with mapping.Document objects.

Parameters:documents – a sequence of dictionaries or Document objects, or objects providing a items() method that can be used to convert them to a dictionary Returns:an iterable over the resulting documents Return type:list Since:version 0.2

update_doc(name, docid=None, **options)

Calls server side update handler.

Parameters:

• name – the name of the update handler function in the format designdoc/updatename.
• docid – optional ID of a document to pass to the update handler.
• options – additional (optional) params to pass to the underlying http resource handler, including headers, body, and `path`. Other arguments will be treated as query string params. See couchdb.http.Resource

Returns:

(headers, body) tuple, where headers is a dict of headers returned from the list function and body is a readable file-like instance

view(name, wrapper=None, **options)

Execute a predefined view.

>>> server = Server()
>>> db = server.create('python-tests')
>>> db['gotham'] = dict(type='City', name='Gotham City')

>>> for row in db.view('_all_docs'):
...     print(row.id)
gotham

>>> del server['python-tests']

Parameters:

• name – the name of the view; for custom views, use the format design_docid/viewname, that is, the document ID of the design document and the name of the view, separated by a slash
• wrapper – an optional callable that should be used to wrap the result rows
• options – optional query string parameters

Returns:

the view results

Return type:

ViewResults

3.3. Document

class couchdb.client.Document

Representation of a document in the database.

This is basically just a dictionary with the two additional properties id and rev, which contain the document ID and revision, respectively.

id

The document ID.

Return type:basestring

rev

The document revision.

Return type:basestring

3.4. ViewResults

class couchdb.client.ViewResults(view, options)

Representation of a parameterized view (either permanent or temporary) and the results it produces.

This class allows the specification of key, startkey, and endkey options using Python slice notation.

>>> server = Server()
>>> db = server.create('python-tests')
>>> db['johndoe'] = dict(type='Person', name='John Doe')
>>> db['maryjane'] = dict(type='Person', name='Mary Jane')
>>> db['gotham'] = dict(type='City', name='Gotham City')
>>> map_fun = '''function(doc) {
...     emit([doc.type, doc.name], doc.name);
... }'''
>>> results = db.query(map_fun)

At this point, the view has not actually been accessed yet. It is accessed as soon as it is iterated over, its length is requested, or one of its rows, total_rows, or offset properties are accessed:

>>> len(results)
3

You can use slices to apply startkey and/or endkey options to the view:

>>> people = results[['Person']:['Person','ZZZZ']]
>>> for person in people:
...     print(person.value)
John Doe
Mary Jane
>>> people.total_rows, people.offset
(3, 1)

Use plain indexed notation (without a slice) to apply the key option. Note that as CouchDB makes no claim that keys are unique in a view, this can still return multiple rows:

>>> list(results[['City', 'Gotham City']])
[<Row id=u'gotham', key=[u'City', u'Gotham City'], value=u'Gotham City'>]

>>> del server['python-tests']

offset

The offset of the results from the first row in the view.

This value is 0 for reduce views.

Return type:int

rows

The list of rows returned by the view.

Return type:list

total_rows

The total number of rows in this view.

This value is None for reduce views.

Return type:int or NoneType for reduce views

update_seq

The database update sequence that the view reflects.

The update sequence is included in the view result only when it is explicitly requested using the update_seq=true query option. Otherwise, the value is None.

Return type:int or NoneType depending on the query options

3.5. Row

class couchdb.client.Row

Representation of a row as returned by database views.

doc

The associated document for the row. This is only present when the view was accessed with include_docs=True as a query parameter, otherwise this property will be None.

id

The associated Document ID if it exists. Returns None when it doesn’t (reduce results).