Class: Wgit::Database

Inherits:
Object
  • Object
show all
Includes:
Assertable
Defined in:
lib/wgit/database/database.rb

Overview

Class providing a DB connection and CRUD operations for the Url and Document collections.

Constant Summary collapse

URLS_COLLECTION =

The default name of the urls collection.

:urls
DOCUMENTS_COLLECTION =

The default name of the documents collection.

:documents
TEXT_INDEX =

The default name of the documents collection text search index.

'text_search'
UNIQUE_INDEX =

The default name of the urls and documents collections unique index.

'unique_url'
DEFAULT_TEXT_INDEX =

The documents collection default text search index. Use db.text_index = Wgit::Database::DEFAULT_TEXT_INDEX to revert changes.

{
  title: 2,
  description: 2,
  keywords: 2,
  text: 1
}.freeze

Constants included from Assertable

Assertable::DEFAULT_DUCK_FAIL_MSG, Assertable::DEFAULT_REQUIRED_KEYS_MSG, Assertable::DEFAULT_TYPE_FAIL_MSG, Assertable::NON_ENUMERABLE_MSG

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Assertable

#assert_arr_types, #assert_required_keys, #assert_respond_to, #assert_types

Constructor Details

#initialize(connection_string = nil) ⇒ Database

Initializes a connected database client using the provided connection_string or ENV['WGIT_CONNECTION_STRING'].

Parameters:

  • connection_string (String) (defaults to: nil) —

    The connection string needed to connect to the database.

Raises:

  • (StandardError) —

    If a connection string isn't provided, either as a parameter or via the environment.



58
59
60
61
62
63
64
65
66
 
# File 'lib/wgit/database/database.rb', line 58

def initialize(connection_string = nil)
  connection_string ||= ENV['WGIT_CONNECTION_STRING']
  raise "connection_string and ENV['WGIT_CONNECTION_STRING'] are nil" \
  unless connection_string

  @client = Database.establish_connection(connection_string)
  @connection_string = connection_string
  @text_index = DEFAULT_TEXT_INDEX
end

Instance Attribute Details

#client ⇒ Object (readonly)

The database client object. Gets set when a connection is established.



42
43
44
 
# File 'lib/wgit/database/database.rb', line 42

def client
  @client
end

#connection_string ⇒ Object (readonly)

The connection string for the database.



39
40
41
 
# File 'lib/wgit/database/database.rb', line 39

def connection_string
  @connection_string
end

#last_result ⇒ Object (readonly)

The raw MongoDB result of the most recent operation.



49
50
51
 
# File 'lib/wgit/database/database.rb', line 49

def last_result
  @last_result
end

#text_index ⇒ Object

The documents collection text index, used to search the DB. A custom setter method is also provided for changing the search logic.



46
47
48
 
# File 'lib/wgit/database/database.rb', line 46

def text_index
  @text_index
end

Class Method Details

.connect(connection_string = nil) ⇒ Wgit::Database

A class alias for Database.new.

Parameters:

  • connection_string (String) (defaults to: nil) —

    The connection string needed to connect to the database.

Returns:

Raises:

  • (StandardError) —

    If a connection string isn't provided, either as a parameter or via the environment.



75
76
77
 
# File 'lib/wgit/database/database.rb', line 75

def self.connect(connection_string = nil)
  new(connection_string)
end

.establish_connection(connection_string) ⇒ Mong::Client

Initializes a connected database client using the connection string.

Parameters:

  • connection_string (String) —

    The connection string needed to connect to the database.

Returns:

  • (Mong::Client) —

    The connected MongoDB client.

Raises:

  • (StandardError) —

    If a connection cannot be established.



85
86
87
88
89
90
91
92
93
 
# File 'lib/wgit/database/database.rb', line 85

def self.establish_connection(connection_string)
  # Only log for error (and more severe) scenarios.
  Mongo::Logger.logger          = Wgit.logger.clone
  Mongo::Logger.logger.progname = 'mongo'
  Mongo::Logger.logger.level    = Logger::ERROR

  # Connects to the database here.
  Mongo::Client.new(connection_string)
end

Instance Method Details

#bulk_upsert(objs) ⇒ Integer

Bulk upserts the objects in the database collection. You cannot mix collection objs types, all must be Urls or Documents.

Parameters:

Returns:

  • (Integer) —

    The total number of upserted objects.



201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
 
# File 'lib/wgit/database/database.rb', line 201

def bulk_upsert(objs)
  assert_arr_types(objs, [Wgit::Url, Wgit::Document])
  raise 'objs is empty' if objs.empty?

  collection = nil
  request_objs = objs.map do |obj|
    collection, query, model = get_type_info(obj)
    data_hash = model.merge(Wgit::Model.common_update_data)

    {
      update_many: {
        filter: query,
        update: { '$set' => data_hash },
        upsert: true
      }
    }
  end

  result = @client[collection].bulk_write(request_objs)
  result.upserted_count + result.modified_count
ensure
  @last_result = result
end

#clear_db ⇒ Integer Also known as: clear_db!

Deletes everything in the urls and documents collections. This will nuke the entire database so yeah... be careful.

Returns:

  • (Integer) —

    The number of deleted records.



550
551
552
 
# File 'lib/wgit/database/database.rb', line 550

def clear_db
  clear_urls + clear_docs
end

#clear_docs ⇒ Integer

Deletes everything in the documents collection.

Returns:

  • (Integer) —

    The number of deleted records.



539
540
541
542
543
544
 
# File 'lib/wgit/database/database.rb', line 539

def clear_docs
  result = @client[DOCUMENTS_COLLECTION].delete_many({})
  result.n
ensure
  @last_result = result
end

#clear_urls ⇒ Integer

Deletes everything in the urls collection.

Returns:

  • (Integer) —

    The number of deleted records.



529
530
531
532
533
534
 
# File 'lib/wgit/database/database.rb', line 529

def clear_urls
  result = @client[URLS_COLLECTION].delete_many({})
  result.n
ensure
  @last_result = result
end

#crawled_urls(limit: 0, skip: 0) {|url| ... } ⇒ Array<Wgit::Url>

Returns Url records that have been crawled.

Parameters:

  • limit (Integer) (defaults to: 0) —

    The max number of Url's to return. 0 returns all.

  • skip (Integer) (defaults to: 0) —

    Skip n amount of Url's.

Yields:

  • (url) —

    Given each Url object (Wgit::Url) returned from the DB.

Returns:

  • (Array<Wgit::Url>) —

    The crawled Urls obtained from the DB.



280
281
282
 
# File 'lib/wgit/database/database.rb', line 280

def crawled_urls(limit: 0, skip: 0, &block)
  urls(crawled: true, limit:, skip:, &block)
end

#create_collections ⇒ nil

Creates the 'urls' and 'documents' collections.

Returns:

  • (nil) —

    Always returns nil.



100
101
102
103
104
105
 
# File 'lib/wgit/database/database.rb', line 100

def create_collections
  @client[URLS_COLLECTION].create
  @client[DOCUMENTS_COLLECTION].create

  nil
end

#create_unique_indexes ⇒ nil

Creates the urls and documents unique 'url' indexes.

Returns:

  • (nil) —

    Always returns nil.



110
111
112
113
114
115
116
117
118
119
120
 
# File 'lib/wgit/database/database.rb', line 110

def create_unique_indexes
  @client[URLS_COLLECTION].indexes.create_one(
    { url: 1 }, name: UNIQUE_INDEX, unique: true
  )

  @client[DOCUMENTS_COLLECTION].indexes.create_one(
    { 'url.url' => 1 }, name: UNIQUE_INDEX, unique: true
  )

  nil
end

#delete(obj) ⇒ Integer

Deletes a record from the database with the matching 'url' field. Pass either a Wgit::Url or Wgit::Document instance.

Parameters:

Returns:

  • (Integer) —

    The number of records deleted - should always be 0 or 1 because urls are unique.



518
519
520
521
522
523
524
 
# File 'lib/wgit/database/database.rb', line 518

def delete(obj)
  collection, query = get_type_info(obj)
  result = @client[collection].delete_one(query)
  result.n
ensure
  @last_result = result
end

#doc?(doc) ⇒ Boolean

Returns whether or not a record with the given doc 'url.url' field (which is unique) exists in the database's 'documents' collection.

Parameters:

Returns:

  • (Boolean) —

    True if doc exists, otherwise false.



465
466
467
468
469
 
# File 'lib/wgit/database/database.rb', line 465

def doc?(doc)
  assert_type(doc, Wgit::Document)
  query = { 'url.url' => doc.url }
  retrieve(DOCUMENTS_COLLECTION, query, limit: 1).any?
end

#docs(limit: 0, skip: 0) {|doc| ... } ⇒ Array<Wgit::Document>

Returns all Document records from the DB. Use #search to filter based on the text_index of the collection.

All Documents are sorted by date_added ascending, in other words the first doc returned is the first one that was inserted into the DB.

Parameters:

  • limit (Integer) (defaults to: 0) —

    The max number of returned records. 0 returns all.

  • skip (Integer) (defaults to: 0) —

    Skip n records.

Yields:

  • (doc) —

    Given each Document object (Wgit::Document) returned from the DB.

Returns:



238
239
240
241
242
243
244
245
246
247
248
 
# File 'lib/wgit/database/database.rb', line 238

def docs(limit: 0, skip: 0, &block)
  results = retrieve(DOCUMENTS_COLLECTION, {},
                     sort: { date_added: 1 }, limit:, skip:)
  return [] if results.count < 1 # results#empty? doesn't exist.

  # results.respond_to? :map! is false so we use map and overwrite the var.
  results = results.map { |doc_hash| Wgit::Document.new(doc_hash) }
  results.each(&block) if block_given?

  results
end

#exists?(obj) ⇒ Boolean

Returns if a record exists with the given obj's url.

Parameters:

Returns:

  • (Boolean) —

    True if a record exists with the url, false otherwise.



476
477
478
 
# File 'lib/wgit/database/database.rb', line 476

def exists?(obj)
  obj.is_a?(String) ? url?(obj) : doc?(obj)
end

#get(obj) ⇒ Wgit::Url, ...

Returns a record from the database with the matching 'url' field; or nil. Pass either a Wgit::Url or Wgit::Document instance.

Parameters:

Returns:



486
487
488
489
490
491
492
493
 
# File 'lib/wgit/database/database.rb', line 486

def get(obj)
  collection, query = get_type_info(obj)

  record = retrieve(collection, query, limit: 1).first
  return nil unless record

  obj.class.new(record)
end

#insert(data) ⇒ Object

Insert one or more Url or Document objects into the DB.

Parameters:

Raises:

  • (StandardError) —

    If data isn't valid.



164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
 
# File 'lib/wgit/database/database.rb', line 164

def insert(data)
  collection = nil
  request_obj = nil

  if data.respond_to?(:map)
    request_obj = data.map do |obj|
      collection, _, model = get_type_info(obj)
      model
    end
  else
    collection, _, model = get_type_info(data)
    request_obj = model
  end

  create(collection, request_obj)
end

#num_docs ⇒ Integer

Returns the total number of Document records in the DB.

Returns:

  • (Integer) —

    The current number of Document records.



438
439
440
 
# File 'lib/wgit/database/database.rb', line 438

def num_docs
  @client[DOCUMENTS_COLLECTION].count
end

#num_records ⇒ Integer Also known as: num_objects

Returns the total number of records (urls + docs) in the DB.

Returns:

  • (Integer) —

    The current number of URL and Document records.



445
446
447
 
# File 'lib/wgit/database/database.rb', line 445

def num_records
  num_urls + num_docs
end

#num_urls ⇒ Integer

Returns the total number of URL records in the DB.

Returns:

  • (Integer) —

    The current number of URL records.



431
432
433
 
# File 'lib/wgit/database/database.rb', line 431

def num_urls
  @client[URLS_COLLECTION].count
end

#search(query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0) {|doc| ... } ⇒ Array<Wgit::Document>

Searches the database's Documents for the given query.

The searched fields are decided by the text index setup on the documents collection. Currently we search against the following fields: "author", "keywords", "title" and "text" by default.

The MongoDB search algorithm ranks/sorts the results in order (highest first) based on each document's "textScore" (which records the number of query hits). The "textScore" is then stored in each Document result object for use elsewhere if needed; accessed via Wgit::Document#score.

Parameters:

  • query (String) —

    The text query to search with.

  • case_sensitive (Boolean) (defaults to: false) —

    Whether character case must match.

  • whole_sentence (Boolean) (defaults to: true) —

    Whether multiple words should be searched for separately.

  • limit (Integer) (defaults to: 10) —

    The max number of results to return.

  • skip (Integer) (defaults to: 0) —

    The number of results to skip.

Yields:

  • (doc) —

    Given each search result (Wgit::Document) returned from the DB.

Returns:

  • (Array<Wgit::Document>) —

    The search results obtained from the DB.



314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
 
# File 'lib/wgit/database/database.rb', line 314

def search(
  query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0
)
  query = query.to_s.strip
  query.replace("\"#{query}\"") if whole_sentence

  # Sort based on the most search hits (aka "textScore").
  # We use the sort_proj hash as both a sort and a projection below.
  sort_proj = { score: { :$meta => 'textScore' } }
  query = {
    :$text => {
      :$search => query,
      :$caseSensitive => case_sensitive
    }
  }

  results = retrieve(DOCUMENTS_COLLECTION, query,
                     sort: sort_proj, projection: sort_proj,
                     limit:, skip:)

  results.map do |mongo_doc|
    doc = Wgit::Document.new(mongo_doc)
    yield(doc) if block_given?
    doc
  end
end

#search!(query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0, sentence_limit: 80) {|doc| ... } ⇒ Array<Wgit::Document>

Searches the database's Documents for the given query and then searches each result in turn using doc.search!. This method is therefore the equivalent of calling Wgit::Database#search and then Wgit::Document#search! in turn. See their documentation for more info.

Parameters:

  • query (String) —

    The text query to search with.

  • case_sensitive (Boolean) (defaults to: false) —

    Whether character case must match.

  • whole_sentence (Boolean) (defaults to: true) —

    Whether multiple words should be searched for separately.

  • limit (Integer) (defaults to: 10) —

    The max number of results to return.

  • skip (Integer) (defaults to: 0) —

    The number of results to skip.

  • sentence_limit (Integer) (defaults to: 80) —

    The max length of each search result sentence.

Yields:

  • (doc) —

    Given each search result (Wgit::Document) returned from the DB having called doc.search!(query).

Returns:

  • (Array<Wgit::Document>) —

    The search results obtained from the DB having called doc.search!(query).



358
359
360
361
362
363
364
365
366
367
368
369
370
 
# File 'lib/wgit/database/database.rb', line 358

def search!(
  query, case_sensitive: false, whole_sentence: true,
  limit: 10, skip: 0, sentence_limit: 80
)
  results = search(query, case_sensitive:, whole_sentence:, limit:, skip:)

  results.each do |doc|
    doc.search!(query, case_sensitive:, whole_sentence:, sentence_limit:)
    yield(doc) if block_given?
  end

  results
end

#search_text(query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0, sentence_limit: 80, top_result_only: false) {|doc| ... } ⇒ Hash<String, String | Array<String>>

Searches the database's Documents for the given query and then searches each result in turn using doc.search. Instead of an Array of Documents, this method returns a Hash of the docs url => search_results creating a search engine like result set for quick access to text matches.

Parameters:

  • query (String) —

    The text query to search with.

  • case_sensitive (Boolean) (defaults to: false) —

    Whether character case must match.

  • whole_sentence (Boolean) (defaults to: true) —

    Whether multiple words should be searched for separately.

  • limit (Integer) (defaults to: 10) —

    The max number of results to return.

  • skip (Integer) (defaults to: 0) —

    The number of results to skip.

  • sentence_limit (Integer) (defaults to: 80) —

    The max length of each search result sentence.

  • top_result_only (Boolean) (defaults to: false) —

    Whether to return all of the documents search results or just the top (most relavent) result.

Yields:

  • (doc) —

    Given each search result (Wgit::Document) returned from the DB.

Returns:

  • (Hash<String, String | Array<String>>) —

    The search results obtained from the DB having mapped the docs url => search_results. The format of search_results depends on the value of top_result_only.



392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
 
# File 'lib/wgit/database/database.rb', line 392

def search_text(
  query, case_sensitive: false, whole_sentence: true,
  limit: 10, skip: 0, sentence_limit: 80, top_result_only: false
)
  results = search(query, case_sensitive:, whole_sentence:, limit:, skip:)

  results
    .map do |doc|
      yield(doc) if block_given?

      # Only return result if its text has a match - compact is called below.
      results = doc.search(
        query, case_sensitive:, whole_sentence:, sentence_limit:
      )
      next nil if results.empty?

      [doc.url, (top_result_only ? results.first : results)]
    end
    .compact
    .to_h
end

#size ⇒ Integer

Returns the current size of the database.

Returns:

  • (Integer) —

    The current size of the DB.



424
425
426
 
# File 'lib/wgit/database/database.rb', line 424

def size
  stats[:dataSize]
end

#stats ⇒ BSON::Document#[]#fetch

Returns statistics about the database.

Returns:

  • (BSON::Document#[]#fetch) —

    Similar to a Hash instance.



417
418
419
 
# File 'lib/wgit/database/database.rb', line 417

def stats
  @client.command(dbStats: 0).documents[0]
end

#uncrawled_urls(limit: 0, skip: 0) {|url| ... } ⇒ Array<Wgit::Url>

Returned Url records that haven't yet been crawled.

Parameters:

  • limit (Integer) (defaults to: 0) —

    The max number of Url's to return. 0 returns all.

  • skip (Integer) (defaults to: 0) —

    Skip n amount of Url's.

Yields:

  • (url) —

    Given each Url object (Wgit::Url) returned from the DB.

Returns:

  • (Array<Wgit::Url>) —

    The uncrawled Urls obtained from the DB.



290
291
292
 
# File 'lib/wgit/database/database.rb', line 290

def uncrawled_urls(limit: 0, skip: 0, &block)
  urls(crawled: false, limit:, skip:, &block)
end

#update(obj) ⇒ Integer

Update a Url or Document object in the DB.

Parameters:

Returns:

  • (Integer) —

    The number of updated records/objects.

Raises:

  • (StandardError) —

    If the obj is not valid.



502
503
504
505
506
507
 
# File 'lib/wgit/database/database.rb', line 502

def update(obj)
  collection, query, model = get_type_info(obj)
  data_hash = model.merge(Wgit::Model.common_update_data)

  mutate(collection, query, { '$set' => data_hash })
end

#upsert(obj) ⇒ Boolean

Inserts or updates the object in the database.

Parameters:

Returns:

  • (Boolean) —

    True if inserted, false if updated.



185
186
187
188
189
190
191
192
193
 
# File 'lib/wgit/database/database.rb', line 185

def upsert(obj)
  collection, query, model = get_type_info(obj)
  data_hash = model.merge(Wgit::Model.common_update_data)
  result = @client[collection].replace_one(query, data_hash, upsert: true)

  result.matched_count.zero?
ensure
  @last_result = result
end

#url?(url) ⇒ Boolean

Returns whether or not a record with the given 'url' field (which is unique) exists in the database's 'urls' collection.

Parameters:

  • url (Wgit::Url) —

    The Url to search the DB for.

Returns:

  • (Boolean) —

    True if url exists, otherwise false.



454
455
456
457
458
 
# File 'lib/wgit/database/database.rb', line 454

def url?(url)
  assert_type(url, String) # This includes Wgit::Url's.
  query = { url: }
  retrieve(URLS_COLLECTION, query, limit: 1).any?
end

#urls(crawled: nil, limit: 0, skip: 0) {|url| ... } ⇒ Array<Wgit::Url>

Returns all Url records from the DB.

All Urls are sorted by date_added ascending, in other words the first url returned is the first one that was inserted into the DB.

Parameters:

  • crawled (Boolean) (defaults to: nil) —

    Filter by Url#crawled value. nil returns all.

  • limit (Integer) (defaults to: 0) —

    The max number of Url's to return. 0 returns all.

  • skip (Integer) (defaults to: 0) —

    Skip n amount of Url's.

Yields:

  • (url) —

    Given each Url object (Wgit::Url) returned from the DB.

Returns:

  • (Array<Wgit::Url>) —

    The Urls obtained from the DB.



260
261
262
263
264
265
266
267
268
269
270
271
272
 
# File 'lib/wgit/database/database.rb', line 260

def urls(crawled: nil, limit: 0, skip: 0, &block)
  query = crawled.nil? ? {} : { crawled: }
  sort = { date_added: 1 }

  results = retrieve(URLS_COLLECTION, query, sort:, limit:, skip:)
  return [] if results.count < 1 # results#empty? doesn't exist.

  # results.respond_to? :map! is false so we use map and overwrite the var.
  results = results.map { |url_doc| Wgit::Url.new(url_doc) }
  results.each(&block) if block_given?

  results
end