INNOPS - INNOVATE OPS Patentsearch

Please load "Patentsearch"

For patent searches or more information about this patent tool, please register and/or login first!

INNOPS - INNOVATE and OPS

INNOVATE and OPS

INNOVATE OPS Patentsearch uses the access to the European Patent Office (EPO) and it's Open Patent Services OPS.
The European Patent Office has a database with detailed descriptions of more then 120 million patents, concentrating technological knowledge from more than 100 countries.

INNOVATE OPS Patentsearch consists of

an implementation of the OPS API Interface for accessing the OPS.

a Solr based archiving server to save results and expand search capabilities

some alternatively implemented or borrowed tools as described on Patent2Net

an UI for OPS and Archive Queries for patent list and details display

an open API to OPS and the Archiving Server with consistent results

INNOVATE OPS Patentsearch is dedicated to:

enable patent information search

provide some statistical analysis and representations of a set of patents

deliver a working model for using OPS data

provide a development base for further OPS patent research client development

Highlights

up to 2000 results per query from OPS.

archive queries:

have no practical limit for query terms
configurable limit for results per query.
may search for data in any used language
all OPS query results can be archived per configuration
or by query utilities
results are prepared to be analyzed immediately
named analysis saved as web pages
interactive pivot table grid and chart analysis
more than 11 Mio documents allready archived.
all EP documents, ca. 7 Mio, regularly updated
open API, not bound to this server's client

Caveats

Query syntax for OPS differs from archive query syntax

About INNOPS

INNOPS - INNOVATE and EPO Open Patent Services

The INNOVATE OPS API was developed during the redesign and new development of a Patent Monitoring System. Instead of the expensable and manual collecting of patents into a database before, this API enabled online searching and saving of patent data according to any definable profile. Although OPS doesn't offer complete information for every patent, at first it seemed to deliver enough information for doing the job the Patent Monitoring System standed for (e.g.: checking search profiles and inform about for new patents or patents changes).

Working with OPS however, we saw some crucial differences between searching in a database with full publication data and OPS:

Until now, in OPS it's not possible to search for text not indexed as English.
This means, that a document with only an other language than English cannot be found per CQL-search, e.g.:
- "Datos auxiliares de profundidad" should deliver ES2645639
- "Das Wärmepaneel" should deliver at least DE202018001836
- "MATERIAU POLYCRISTALLIN ET SON PROCEDE D'OBTENTION" should deliver at least FR3066952
- "Inrichting voor het lanceren van dummy's" should deliver NL1042680B1
However, if a text was indexed wrong, you may get results, e.g.: the document EP1223398A1 you may find with "Erwärmen", since the german title was indexed as english.
Single search results could have about 100 documents only.
Some changes in documents couldn't be found, e.g. changes in a B-document, if the publicationdate of an A-document didn't change.
- "pn:EP0000010 AND pd:19800109" should deliver EP0000010B1.
Full text searches in descriptions are not possible, since only biblio data are searched.
In OPS, there is no unique index for a single exchange-doc to search for.
OPS only may have max. 10 terms within one query

INNOVATE (we) decided to work on reducing these limitations and the data preparation for easy analysis. Therefore:

The EPO DOCDB download possibility was analyzed and prepared to use for an archive base. It was discarded by INNOVATE's customer.
With that solution prepared, INNOVATE developed the saving of every query in a SOLR search engine. It delivers it's query results compatable with the OPS back-end.
For data analysis we came across the solutions of Patent2Net (P2N). Unfortunately this system is a bit hard to use, especially for non programmers. Very good however, because of it's results. It inspired us to combine our and some of their possibilities.

INNOPS - Patent Tool Development Status

In the actual development status of INNOVATE OPS Patentsearch we implemented solutions for most of the above mentioned crucial limitations to OPS on a Linux system. (Working with profiles, after cyclic updates, users may be advised to use the Solr archive for searches.):

Query for Results

Maximum ca. 2000 result documents per OPS and Archive queries. Archive csv-downloads tested with 30000 results/query.
Every OPS search result can be saved in a Solr search engine database. (biblio-data)
Practically no limits on query terms per archive query.
Within saved data, UTF-8 searches for any kind of data and within any language are possible.
Continuous saving already prepared more then 11 Mio. documents for full search now, whereof almost all EP patent documents.

Patent Detail

Selectable from results list
Detail info, depending on availability for
* Bibligraphic Data
* Description
* Claims
* Legals
* Citations
* Agent
* Original Document
* Family Drawings

Features

Pivot data and analysis are available immediately after querying OPS or the Solr Archive
Prepared for automatic updates of archive data.
A help menu provides full text IPC/CPC definition queries
Queries are saved in and may be reused from a query list.
csv export for further analysis from query results or, unlimited, from Solr archive
Monitoring of saved query profiles (weekly/monthly)
Prepared for more users and roles (work in progress).
Optimal throttling control according the Fair Use Policy of EPO/OPS.
API for access to main OPS functions.

The actual software therefore is a good basis to start working with patent searches but also for those software developers, interested in building a more complete Patent Search Application.

We are developing on quite a fast Linux Server (i7 / 32 GB RAM) but the base system also runs on a Raspberry Pi. Not fast for some functionality, but working very well!

More Information Patentsearch

For more information you're invited to register yourself on this site. Then you'll find more information and you can test the available system

Example Screens

Query result list

Detail info view

The following graphs are generated immediately after results display:

Interactive tool: heatmap

Interactive tool: line chart

Mapping: Country Applicants

Work in Progress

IPC/CPC Structure Search

Query Database

Here you may look for text within all english IPC/CPC definitions.
Just put terms you want to find within the results into the first input field and put terms you don't want to find within the results into the second field.
All terms in the input fields will be wildcarded and concatinated by "AND".
In the second input field terms also will also be negated by "NOT".
Both terms will be concatinated by "AND", so e.g.:
1. term: "working soil"
2. term: "HUMAN NECESSITIES"
will search for:
*working*" AND "*soil* AND NOT *HUMAN* AND NOT *NECESSITIES*
Results are limited to max. 100 per query.
The underlying database will be updated, wenever you rollover an IPC/CPC term in a query result list.

Query OPS

OPS also provides an IPC/CPC search. To use this, please input your OPS query term in the first input field and select "Search OPS" to get your results from OPS.
The OPS search basically uses the same syntax as the patent search but is restricted to titles and abstracts only. Results are limited to max. 100 per query.

The query result may be ordered by "Percentage" which shows the relevance of each result item.
Rolling over with a mouse, symbols in the result text show actual definitions from OPS.

Using a CPC-Symbol as term for search, the result shows the definitions of the CPC term and it ancestors.

Timerange Query

Ops-Queryterm

Solr-Queryterm (if different)

From Date

To Date