python solr客户端

原文：http://blog.chenlb.com/2010/03/use-solr-python-client.html

 

最近开始慢慢使用 python 了，用 python 处理小事情还是很方便的（比 java 开发效率高）。拿 python 来做 solr 索引也方便。

官方有相关的 solr python 客户端 ，我简单对比了下选择 solrpy ，理由是使用说明详细一点、功能丰富一点、项目感得比较活。

tips：从 solr 1.3 开始官方已经不在维护 solr 客户端了。

安装 solrpy：

easy_install solrpy

# 或者用源码安装

python setup.py install

安装 easy_install 我就不写了，使用 python 的人都知道。

查询示例：

    import solr  
      
    s = solr.SolrConnection('http://127.0.0.1:8080/solr')  
      
    result = s.query("*:*", rows=5)  
    for r in result:  
        print r['id'], r['title']  

 

方便的查询方法。返回的结果用字典类型（如果 object 类型就更方便了）。

官方的示例：

  import solr   
      
    # create a connection to a solr server  
    s = solr.SolrConnection('http://example.org:8083/solr')   
      
    # add a document to the index  
    s.add(id=1, title='Lucene in Action', author=['Erik Hatcher', 'Otis Gospodnetić'])  
    s.commit()   
      
    # do a search  
    response = s.query('title:lucene')  
    for hit in response.results:  
        print hit['title']  

 

方法有 query (q, fields=None, highlight=None, score=True, sort=None, **params) 、add_many(list) 、delete(id) 、delete_query(query)、commit(wait_flush=True, wait_searcher=True)、optimize(wait_flush=True, wait_searcher=True) 等方法，还有一个方便的就是翻页 next_batch()。

 

更多使用请看：http://code.google.com/p/solrpy/source/browse/trunk/solr/core.py 的注释。

越来越喜欢 python 了，python 周边工具/库还是比较多，很容易找到对应的库。当然 solrj 也方便。

 

更多：http://wiki.apache.org/solr/SolPython

 

 

 

A simple Solr client for python.


Features
--------
 * Supports Solr 1.2+
 * Supports http/https and SSL client-side certificates
 * Uses persistent HTTP connections by default
 * Properly converts to/from Solr data types, including datetime objects
 * Supports both querying and update commands (add, delete)
 * Requires Python 2.3+


Connections
-----------
`SolrConnection` can be passed in the following parameters.
Only `url` is required,.

    url -- URI pointing to the Solr instance. Examples:

        http://localhost:8080/solr
        https://solr-server/solr

        Your python install must be compiled with SSL support for the
        https:// schemes to work. (Most pre-packaged pythons are.)

    persistent -- Keep a persistent HTTP connection open.
        Defaults to true.

    timeout -- Timeout, in seconds, for the server to response.
        By default, use the python default timeout (of none?)
        NOTE: This changes the python-wide timeout.

    ssl_key, ssl_cert -- If using client-side key files for
        SSL authentication,  these should be, respectively,
        your PEM key file and certificate file
       
    http_user, http_pass -- If given, include HTTP Basic authentication
        in all request headers.

Once created, a connection object has the following public methods:

    query(q, fields=None, highlight=None,
          score=True, sort=None, **params)

            q -- the query string.

            fields -- optional list of fields to include. It can be either
                a string in the format that Solr expects ('id,f1,f2'), or
                a python list/tuple of field names.   Defaults to returning
                all fields. ("*")

            score -- boolean indicating whether "score" should be included
                in the field list.  Note that if you explicitly list
                "score" in your fields value, then this parameter is
                effectively ignored.  Defaults to true.

            highlight -- indicates whether highlighting should be included.
                `highlight` can either be `False`, indicating "No" (the
                default),  `True`, incidating to highlight any fields
                included in "fields", or a list of field names.

            sort -- list of fields to sort by.

            Any parameters available to Solr 'select' calls can also be
            passed in as named parameters (e.g., fq='...', rows=20, etc).

            Many Solr parameters are in a dotted notation (e.g.,
            `hl.simple.post`).  For such parameters, replace the dots with
            underscores when calling this method. (e.g.,
            hl_simple_post='</pre'>)

            Returns a Response object

    add(**params)

            Add a document.  Pass in all document fields as
            keyword parameters:

                add(id='foo', notes='bar')

            You must "commit" for the addition to be saved.

    add_many(lst)

            Add a series of documents at once.  Pass in a list of
            dictionaries, where each dictionary is a mapping of document
            fields:

                add_many( [ {'id': 'foo1', 'notes': 'foo'},
                            {'id': 'foo2', 'notes': 'w00t'} ] )

            You must "commit" for the addition to be saved.

    delete(id)

            Delete a document by id.

            You must "commit" for the deletion to be saved.

    delete_many(lst)

            Delete a series of documents.  Pass in a list of ids.

            You must "commit" for the deletion to be saved.

    delete_query(query)

            Delete any documents returned by issuing a query.

            You must "commit" for the deletion to be saved.


    commit(wait_flush=True, wait_searcher=True)

            Issue a commit command.

    optimize(wait_flush=True, wait_searcher=True)

            Issue an optimize command.

    raw_query(**params)

            Send a query command (unprocessed by this library) to
            the Solr server. The resulting text is returned un-parsed.

                raw_query(q='id:1', wt='python', indent='on')

            Many Solr parameters are in a dotted notation (e.g.,
            `hl.simple.post`).  For such parameters, replace the dots with
            underscores when calling this method. (e.g.,
            hl_simple_post='</pre'>)

    close()
            Close the underlying HTTP(S) connection.


Query Responses
---------------

    Calls to connection.query() return a Response object.

    Response objects always have the following properties:

        results -- A list of matching documents. Each document will be a
            dict of field values.

        results.start -- An integer indicating the starting # of documents

        results.numFound -- An integer indicating the total # of matches.

        results.maxScore -- An integer indicating the maximum score assigned
                            to a document. Takes into account all of documents
                            found by the query, not only the current batch.

        header -- A dict containing any responseHeaders.  Usually:

            header['params'] -- dictionary of original parameters used to
                        create this response set.

            header['QTime'] -- time spent on the query

            header['status'] -- status code.

            See Solr documentation for other/typical return values.
            This may be settable at the Solr-level in your config files.


        next_batch() -- If only a partial set of matches were returned
            (by default, 10 documents at a time), then calling
            .next_batch() will return a new Response object containing
            the next set of matching documents. Returns None if no
            more matches.

            This works by re-issuing the same query to the backend server,
            with a new 'start' value.

        previous_batch() -- Same as next_batch, but return the previous
            set of matches.  Returns None if this is the first batch.

    Response objects also support __len__ and iteration. So, the following
    shortcuts work:

        responses = connection.query('q=foo')
        print len(responses)
        for document in responses:
            print document['id'], document['score']


    If you pass in `highlight` to the SolrConnection.query call,
    then the response object will also have a "highlighting" property,
    which will be a dictionary.


Quick examples on use:
----------------------

Example showing basic connection/transactions

    >>> from solr import *
    >>> c = SolrConnection('http://localhost:8983/solr')
    >>> c.add(id='500', name='python test doc', inStock=True)
    >>> c.delete('123')
    >>> c.commit()


Examples showing the search wrapper

    >>> response = c.query('test', rows=20)
    >>> print response.results.start
     0
    >>> for match in response:
    ...     print match['id'],
      0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
    >>> response = response.next_batch()
    >>> print response.results.start
     20

Enter a raw query, without processing the returned HTML contents.

    >>> print c.raw_query(q='id:[* TO *]', wt='python', rows='10')