Third party search applications for a search system

• US 10,019,484 B2
• Filed: 08/06/2013
• Issued: 07/10/2018
• Est. Priority Date: 07/17/2013
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Fink, Michael, Efron, Niv, Fink, Eyal, Gontmakher, Alex, Vorobey, Anatoly, Matias, Yossi, Menzel, Jack Wright, Arnon, Asaph
• Primary Examiner(s): Bibbee, Jared M
• Application Number:
  US13/960,302
• Publication Number:
  US 20150026212A1
• Time in Patent Office: 1,799 Days
• CPC Class Codes: G06F 16/24553 of query operations
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 21/53 by executing in a restricted environment, e.g. sandbox or secure virtual machine
• CPC Class Codes: G06F 21/74 operating in dual or compartmented mode, i.e. at least one secure mode

Abstract

Systems and methods offer a search system with third-party provided search applications that are triggered in response to specified queries and run at the search system. For example, a method may include determining that a query triggers a third party search application hosted at the search system, extracting a parameter from the query based on a query template, executing the third party search application with the parameter in a sandboxed manner at the search system to generate a third-party formatted answer for the query, and providing the third-party formatted answer as a search result for the query. The third party may provide the query template, parameter attributes, and the third party formatted answer. The third party search application may be stored at the search system and include the query template, a data store, the parameter attributes, and instructions for accessing the data store using the parameter.

Claims

1. A computer-implemented method comprising:

generating, using at least one processor of a search system, first answers for a query from an inverted index for a document corpus;

determining, using the at least one processor, that the query corresponds to a query template, the query template being a regular expression with a variable portion that corresponds to a parameter;

extracting a parameter value from the query that corresponds to the portion of the query that matches the parameter;

generating, using the at least one processor of the search system, a third-party formatted natural language answer for the query by executing computer-language instructions obtained from the third party for using the parameter value to generate the natural language answer and for formatting the natural language answer, the executing occurring in a sandboxed environment; and

providing the third-party formatted natural language answer with the first answers as a search result for the query,

wherein the query template and the computer-language instructions are stored at the search system as a third party search application, which extracts the parameter value and generates the third party formatted natural language answer without communicating with a third party site.

2. The method of claim 1, wherein the third party search application includes a data store received from the third party and stored at the search system and the computer-language instructions include instructions for accessing the data store using the parameter.

3. The method of claim 2, wherein the third party search application generates the third-party formatted natural language answer by:

using the parameter value to look up a record in the data store; and

using a field from the record to replace a second parameter in the third-party formatted answer.

4. The method of claim 2, further comprising:

receiving updated data for the data store from the third party on a periodic basis; and

updating the data store.

5. The method of claim 2, further including receiving, at the search system, the query template, information describing the parameter, and information describing the data store from the third party.

6. The method of claim 2, wherein the instructions for accessing the data store are selected by the third party from a predefined set of instructions.

7. The method of claim 1, wherein the third-party formatted natural language answer consists of a sentence determined by the third party that includes the parameter value.

8. The method of claim 1, further comprising:

repeating the generating, determining, extracting, and providing over multiple queries;

monitoring a quantity of times the first answers are selected; and

responsive to determining that the quantity of times meets a threshold, disabling the query template.

9. The method of claim 1, wherein determining that the query corresponds to the query template includes:

determining a source for the query template; and

determining that signals for the source meet a threshold before determining that the query corresponds to the query template.

10. The method of claim 1, wherein the first answers are generated concurrently with the third-party formatted natural language answer.

11. A search system comprising:

at least one processor; and

memory storing instructions that, when executed by the at least one processor, cause the search system to perform operations comprising;

receiving a query template from a third-party, the query template comprising a regular expression with at least one parameter and receiving one or more attributes for the parameter,

receiving a data store from the third party, the data store including attributes and data,

receiving, from the third party computer-language instructions for accessing the data store using the parameter and for formatting a natural-language query answer that includes data identified by accessing the data store, and

storing the data for data store, the attributes of the data store, the parameter attributes, and the query template in sandboxed memory at the search system as a third party search application,

wherein the search system executes the third party search application, which generates and provides the natural-language query answer responsive to a query matching the query without communicating with a third party site.

12. The search system of claim 11, wherein the receiving and storing are repeated for a plurality of distinct third parties.

13. The search system of claim 11, wherein the memory further stores instructions that, when executed by the at least one processor, cause the search system to perform operations comprising:

determining that the query corresponds to the query template;

generating a third-party formatted answer for the query by providing a parameter value extracted from a portion of the query that matches the parameter to the third party search application and executing the instructions for accessing the data store; and

providing the third-party formatted answer as a search result for the query.

14. The search system of claim 13, further comprising an inverted index of documents in a document corpus and wherein the memory further stores instructions that, when executed by the at least one processor, cause the search system perform operations comprising:

generating snippet-based answers for the query by searching the index of documents responsive to the query; and

providing the snippet-based answers and the third-party formatted answer as the search result for the query.

15. The search system of claim 11, further comprising:

a graph-based data store storing entities linked by relationships,

wherein the attributes for the parameter include an entity collection,

and wherein generating the natural language query answer includes;

determining that the query corresponds to the query template,

determining a portion of the query that corresponds to the parameter of the query template,

determining that the portion of the query that corresponds the parameter corresponds to an attribute of an entity in the graph-based data store,

generating the natural language query answer for the query by providing the attribute of the entity to the third party search application and executing the instructions for accessing the data store using the attribute of the entity as the parameter, and

providing the natural language query answer as a search result for the query.

16. The search system of claim 13, wherein the third party search application is one of a plurality of third party search applications hosted at the search system and generating the third-party formatted answer includes:

evaluating query templates associated with the plurality of third party search applications to determine query templates that correspond to the query;

generating respective third-party formatted answers for the query templates that correspond to the query; and

providing the respective third-party formatted answers as part of the search result.

17. The search system of claim 11, wherein the instructions for accessing the data store are limited to a predefined set of instructions.

18. The search system of claim 17, wherein the search system integrates the instructions and the data store with a query engine autonomously of a human operator of the search system.

19. A search system comprising:

at least one processor; and

memory storing instructions that, when executed by the at least one processor, cause the search system to perform operations comprising;

determining that a query triggers a third party search application hosted at the search system based on a query template, the query template being a regular expression with a variable portion and a fixed portion, and wherein the query triggers the third party search application responsive to determining that the query includes terms that match the fixed portion,

extracting a parameter value from the query using the query template, wherein the parameter value corresponds to the portion of the query that matches the variable portion,

executing computer-language instructions obtained from the third party for using the parameter value to generate a third-party formatted natural language answer for the query and for formatting the natural language answer, the executing occurring in a sandboxed manner at the search system, and

providing the third-party formatted natural language answer as a search result for the query,

wherein the query template and the computer-language instructions are stored at the search system as the third party search application, which extracts the parameter value and generates the third-party formatted natural language answer without communicating with a third party site.

20. The search system of claim 19, the memory further storing instructions that, when executed by the at least one processor, cause the search system to provide, with the third-party formatted natural language answer, a link that, responsive to being selected by a query requester, indicates disapproval of the third-party formatted natural language answer.

21. The search system of claim 19, wherein determining that the query triggers the third party search application includes:

determining that the query includes a pronoun; and

substituting the pronoun in the query with a noun from a previous query submitted by a query requester of the query.

22. The search system of claim 19, wherein the memory also stores instructions that, when executed by the at least one processor, causes the search system to generate first answers for the query from an inverted index and provide the first answers with the third-party formatted natural language answer as the search result.

23. The search system of claim 22, wherein determining that the query triggers the third party search application includes determining that a website associated with the third party search application appears in the first answers.

24. The search system of claim 23, wherein determining that the website appears in the first answers includes determining that a ranking signal for the website meets a signal threshold.