Class: Ferret::Index::FieldInfos

Summary

The FieldInfos class holds all the field descriptors for an index. It is this class that is used to create a new index using the FieldInfos#create_index method. If you are happy with the default properties for FieldInfo then you don‘t need to worry about this class. IndexWriter can create the index for you. Otherwise you should set up the index like in the example;

Example

  field_infos = FieldInfos.new(:term_vector => :no)

  field_infos.add_field(:title, :index => :untokenized, :term_vector => :no,
                        :boost => 10.0)

  field_infos.add_field(:content)

  field_infos.add_field(:created_on, :index => :untokenized_omit_norms,
                        :term_vector => :no)

  field_infos.add_field(:image, :store => :compressed, :index => :no,
                        :term_vector => :no)

  field_infos.create_index("/path/to/index")

Default Properties

See FieldInfo for the available field property values.

When you create the FieldInfos object you specify the default properties for the fields. Often you‘ll specify all of the fields in the index before you create the index so the default values won‘t come into play. However, it is possible to continue to dynamically add fields as indexing goes along. If you add a document to the index which has fields that the index doesn‘t know about then the default properties are used for the new field.

Public Class Methods

load (yaml_str)

Load FieldInfos from a YAML file. The YAML file should look something like this: default: 

  store: :yes
  index: :yes
  term_vector: :no

fields: 

  id:
    index: :untokenized
    term_vector: :no

  title:
    boost: 20.0
    term_vector: :no

  content:
    term_vector: :with_positions_offsets

    # File lib/ferret/field_infos.rb, line 24
24:     def self.load(yaml_str)
25:       info = YAML.load(yaml_str)
26:       convert_strings_to_symbols(info)
27:       fis = FieldInfos.new(info[:default])
28:       fields = info[:fields]
29:       fields.keys.each {|key| fis.add_field(key, fields[key])} if fields
30:       fis
31:     end

FieldInfos.new(defaults = {}) → field_infos

Create a new FieldInfos object which uses the default values for fields specified in the default hash parameter. See FieldInfo for available property values. 

/*
 *  call-seq:
 *     FieldInfos.new(defaults = {}) -> field_infos
 *     
 *  Create a new FieldInfos object which uses the default values for fields
 *  specified in the +default+ hash parameter. See FieldInfo for available
 *  property values.
 */
static VALUE
frt_fis_init(int argc, VALUE *argv, VALUE self)
{
    VALUE roptions;
    FieldInfos *fis;
    enum StoreValues store = STORE_YES;
    enum IndexValues index = INDEX_YES;
    enum TermVectorValues term_vector = TERM_VECTOR_WITH_POSITIONS_OFFSETS;
    float boost;

    rb_scan_args(argc, argv, "01", &roptions);
    if (argc > 0) {
        frt_fi_get_params(roptions, &store, &index, &term_vector, &boost);
    }
    fis = fis_new(store, index, term_vector);
    Frt_Wrap_Struct(self, &frt_fis_mark, &frt_fis_free, fis);
    object_add(fis, self);
    return self;
}

Public Instance Methods

fis << fi → fis
fis.add(fi) → fis

Add a FieldInfo object. Use the FieldInfos#add_field method where possible. 

/*
 *  call-seq:
 *     fis << fi -> fis
 *     fis.add(fi) -> fis
 *
 *  Add a FieldInfo object. Use the FieldInfos#add_field method where
 *  possible.
 */
static VALUE
frt_fis_add(VALUE self, VALUE rfi)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    FieldInfo *fi = (FieldInfo *)frt_rb_data_ptr(rfi);
    fis_add_field(fis, fi);
    REF(fi);
    return self;
}

fis[name] → field_info
fis[number] → field_info

Get the FieldInfo object. FieldInfo objects can be referenced by either their field-number of the field-name (which must be a symbol). For example; 

  fi = fis[:name]
  fi = fis[2]

/*
 *  call-seq:
 *     fis[name] -> field_info
 *     fis[number] -> field_info
 *
 *  Get the FieldInfo object. FieldInfo objects can be referenced by either
 *  their field-number of the field-name (which must be a symbol). For
 *  example;
 *
 *    fi = fis[:name]
 *    fi = fis[2]
 */
static VALUE
frt_fis_get(VALUE self, VALUE ridx)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    VALUE rfi = Qnil;
    switch (TYPE(ridx)) {
        case T_FIXNUM: {
            int index = FIX2INT(ridx);
            if (index < 0) index += fis->size;
            if (index < 0 || index >= fis->size) {
                rb_raise(rb_eArgError, "index of %d is out of range (0..%d)\n",
                         index, fis->size);
            }
            rfi = frt_get_field_info(fis->fields[index]);
            break;
                       }
        case T_SYMBOL:
            rfi = frt_get_field_info(fis_get_field(fis, frt_field(ridx)));
            break;
        case T_STRING:
            rfi = frt_get_field_info(fis_get_field(fis, StringValuePtr(ridx)));
            break;
        default:
            rb_raise(rb_eArgError, "Can't index FieldInfos with %s",
                     rs2s(rb_obj_as_string(ridx)));
            break;
    }
    return rfi;
}

fis << fi → fis
fis.add(fi) → fis

Add a FieldInfo object. Use the FieldInfos#add_field method where possible. 

/*
 *  call-seq:
 *     fis << fi -> fis
 *     fis.add(fi) -> fis
 *
 *  Add a FieldInfo object. Use the FieldInfos#add_field method where
 *  possible.
 */
static VALUE
frt_fis_add(VALUE self, VALUE rfi)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    FieldInfo *fi = (FieldInfo *)frt_rb_data_ptr(rfi);
    fis_add_field(fis, fi);
    REF(fi);
    return self;
}

fis.add_field(name, properties = {} → fis

Add a new field to the FieldInfos object. See FieldInfo for a description of the available properties. 

/*
 *  call-seq:
 *     fis.add_field(name, properties = {} -> fis
 *
 *  Add a new field to the FieldInfos object. See FieldInfo for a description
 *  of the available properties.
 */
static VALUE
frt_fis_add_field(int argc, VALUE *argv, VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    FieldInfo *fi;
    enum StoreValues store = fis->store;
    enum IndexValues index = fis->index;
    enum TermVectorValues term_vector = fis->term_vector;
    float boost = 1.0f;
    VALUE rname, roptions;

    rb_scan_args(argc, argv, "11", &rname, &roptions);
    if (argc > 1) {
        frt_fi_get_params(roptions, &store, &index, &term_vector, &boost);
    }
    fi = fi_new(frt_field(rname), store, index, term_vector);
    fi->boost = boost;
    fis_add_field(fis, fi);
    return self;
}

fis.create_index(dir) → self

Create a new index in the directory specified. The directory dir can either be a string path representing a directory on the file-system or an actual directory object. Care should be taken when using this method. Any existing index (or other files for that matter) will be deleted from the directory and overwritten by the new index. 

/*
 *  call-seq:
 *     fis.create_index(dir) -> self
 *     
 *  Create a new index in the directory specified. The directory +dir+ can
 *  either be a string path representing a directory on the file-system or an
 *  actual directory object. Care should be taken when using this method. Any
 *  existing index (or other files for that matter) will be deleted from the
 *  directory and overwritten by the new index.
 */
static VALUE
frt_fis_create_index(VALUE self, VALUE rdir)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    Store *store = NULL;
    if (TYPE(rdir) == T_DATA) {
        store = DATA_PTR(rdir);
        REF(store);
    } else {
        StringValue(rdir);
        frt_create_dir(rdir);
        store = open_fs_store(rs2s(rdir));
    }
    index_create(store, fis);
    store_deref(store);
    return self;
}

fis.each {|fi| do_something } → fis

Iterate through the FieldInfo objects. 

/*
 *  call-seq:
 *     fis.each {|fi| do_something } -> fis
 *
 *  Iterate through the FieldInfo objects.
 */
static VALUE
frt_fis_each(VALUE self)
{
    int i;
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);

    for (i = 0; i < fis->size; i++) {
        rb_yield(frt_get_field_info(fis->fields[i]));
    }
    return self;
}

fis.fields → symbol array

Return a list of the field names (as symbols) of all the fields in the index. 

/*
 *  call-seq:
 *     fis.fields -> symbol array
 *
 *  Return a list of the field names (as symbols) of all the fields in the
 *  index.
 */
static VALUE
frt_fis_get_fields(VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    VALUE rfield_names = rb_ary_new();
    int i;
    for (i = 0; i < fis->size; i++) {
        rb_ary_push(rfield_names, ID2SYM(rb_intern(fis->fields[i]->name)));
    }
    return rfield_names;
}

fis.size → int

Return the number of fields in the FieldInfos object. 

/*
 *  call-seq:
 *     fis.size -> int
 *     
 *  Return the number of fields in the FieldInfos object.
 */
static VALUE
frt_fis_size(VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    return INT2FIX(fis->size);
}

fis.to_a → array

Return an array of the FieldInfo objects contained but this FieldInfos object. 

/*
 *  call-seq:
 *     fis.to_a -> array
 *
 *  Return an array of the FieldInfo objects contained but this FieldInfos
 *  object.
 */
static VALUE
frt_fis_to_a(VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    VALUE rary = rb_ary_new();
    int i;

    for (i = 0; i < fis->size; i++) {
        rb_ary_push(rary, frt_get_field_info(fis->fields[i]));
    }
    return rary;
}

fis.to_s → string

Return a string representation of the FieldInfos object. 

/*
 *  call-seq:
 *     fis.to_s -> string
 *     
 *  Return a string representation of the FieldInfos object.
 */
static VALUE
frt_fis_to_s(VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    char *fis_s = fis_to_s(fis);
    VALUE rfis_s = rb_str_new2(fis_s);
    free(fis_s);
    return rfis_s;
}

fis.tokenized_fields → symbol array

Return a list of the field names (as symbols) of all the tokenized fields in the index. 

/*
 *  call-seq:
 *     fis.tokenized_fields -> symbol array
 *
 *  Return a list of the field names (as symbols) of all the tokenized fields
 *  in the index.
 */
static VALUE
frt_fis_get_tk_fields(VALUE self)
{
    FieldInfos *fis = (FieldInfos *)DATA_PTR(self);
    VALUE rfield_names = rb_ary_new();
    int i;
    for (i = 0; i < fis->size; i++) {
        if (!fi_is_tokenized(fis->fields[i])) continue;
        rb_ary_push(rfield_names, ID2SYM(rb_intern(fis->fields[i]->name)));
    }
    return rfield_names;
}