I use OCR function (DOCUMENT_TEXT_DETECTION) of Google Vision API to process different medical documents. There are some tables in them. According to Google's documentation there is a special BlockType for tables (https://cloud.google.com/vision/docs/reference/rpc/google.cloud.vision.v1#blocktype), however, I can not get this sort of block in a response from Google even in a case when existence of a table is obvious. Does anybody know the reason of such behavior? Do I need to use special options or methods to detect tables using Google Vision API? An example of a table I tried to process:
I found out your question about tables in Google Vision API in Google Forum. The short answer: tables (as blockType) aren't supported now (10/21/2021) but there is a feature request with minor priority: Google Vision API Issue Tracker
I would recommend you to use Document AI: Document AI. I checked and it returned meta info about tables.
Related
When I Google for something and click on a heading from the Knowledge Graph, Google displays a list of the results just below the search bar:
Screenshot here
I typed in the same thing in my Knowledge Graph API:
https://kgsearch.googleapis.com/v1/entities:search?query=apple%20products&key=[MY_KEY]&indent=True&limit=100
and the results are not the same. Short of scraping, is there a way of obtaining the exact list shown within the results via a Google API? It doesn't seem to be Knowledge Graph, or at least the queries are not the same.
Short of scraping, is there a way of obtaining the exact list shown
within the results via a Google API? It doesn't seem to be Knowledge
Graph
Apparently not, you are using the correct API but unfortunately that is how the API works—it usually gives you a different result than if you do a simple Google search.
Why this is the case—I don't know. I decided scrape google instead.
The Knowledge Graph API doesn't have access to the same data sources as a simple Google search. You can read about here if interested (https://web.archive.org/web/20130329151128/http://zecblog.com/2012/09/16/the-short-life-of-the-open-knowledge-graph/).
But yes in short, you'll have to do your own scraping if you want the same information, although it's difficult because Google purposely obfuscates the HTML tags.
Im looking to use Google Adwords Keyword tool data on a website. Ive been looking around in the API and I cant find much to match what I need. I noticed a lot of keyword research tool websites use google as their main source for their information. How would I go about doing this and extracting the data and have it run on a website automatically so it wouldnt need to be updated manually each month?
you can use the Traffic Estimator service in the AdWords API:
https://developers.google.com/adwords/api/docs/reference/v201409/TrafficEstimatorService
Be warned that this is notoriously inaccurate (which is odd given that you would think Google had its own data to call upon!)
I use the TargetingIdea service in the AdWords API to generate lists of keywords to use for building AdWords campaigns. (https://developers.google.com/adwords/api/docs/reference/v201409/TargetingIdeaService.TargetingIdea)
First off you need an API key - they're not that easy to get and your app needs to offer a whole lot of features to meet the required minimum functionality - take a look here https://developers.google.com/adwords/api/docs/requirements
Once you've jumped over that hurdle you get the data from Google by sending a request to the service. That request includes some targeting criteria like location and language and also a "seed" keyword. You can also specify if you want closely related results or broadly related results.
For example if you sold tractors you'd put 'tractors' in as a seed keyword and then the API would return either closely related terms like 'tractors for sale', 'used tractor spares' etc or more broadly related terms like 'agricultural machinery'.
We're creating a custom table in SAP comprising all of the information we need and the customer needs the report from this table uploaded to Google Docs. We do not use Business By Design. Is there any other quicka nd easy way to upload our report?
I don't know much about SAP but the Documents List API has methods to programmatically upload a document to Google Docs: https://developers.google.com/google-apps/documents-list/.
For instance, if you can export the SAP table as a csv file, that can be automatically converted into a Google Spreadsheet during the upload process.
You could also go with a no-programming required solution and install the Google Drive app on a machine with access to the files for automatic sync up to Google Drive:
http://support.google.com/drive/bin/answer.py?hl=en&answer=2374989
suggest you take a look at the SAP Developer Network (SDN) / SAP Community Network (SCN) where there is a project called ABAP2GAPPS that has done this.
Note the ABAP2GAPPS example is a bit difficult to figure out (but you can learn a lot from it), and it also uses the OAuth2 'authorization code flow" OAuth2.0 flow/pattern, which requires an end-user 'consent' in an browser pop-up...so if you want to push up a file from ABAP automatically from a background job without end-user interaction then ABAP2GAPPS is not the full answer (but again, ABAP2GAPPS is a great example, suggest you look at it.)
We recently were able to achieve an interface from SAP ABAP to the Google Fusion Table API using OAuth2, with only about a 100 lines of ABAP...and the techniques we employed could be used on any of the Google API's...here's a link to the video:
Link to YouTube video interface ABAP to Google API
hope you find this helpful
I need to provide search functionality as the cornerstone of my app (basically I think I can make a better interface than whats is offered by Google, bing, yahoo - bold claim I know). Therefore I will not build an entire search engine from scratch, I will rather use an API provider by a major search engine and simply modify the interface that present the results to my visitors. I am a Google fanboi and I initially tried to use Google Search AJAX API. However, I was very disappointed since I can barely change anything at all.
Specifically;
I need an API that will let me pull programmatically the results from a major search engine and let me output them with the style and goodies that I want.
I am not looking to alter the order in which the search results are presented to the user by inserting crap in between the good results (this is against my philosophy). But I would like to wrap massive php/javascript around each search results so that I can completely control how each result is graphically rendered.
I heard of the BING API 2.0, would that be more flexible than Google AJAX API? Could anyone provide output if any of the current search providers are offering API for this purpose right now?
After careful investigation it appears to me that the Bing 2.0 API is the most flexible and robust search API currently on the market.
The Bing API does require you to put their ads next to the results. You can place your own ads with the Yahoo API. The Yahoo API uses the same search technology (same results) and can be used starting from $0,40 per 1000 queries. The lowest, bulk price of the Google API is $4,00 per 1000 queries. Google's relevance is easily ten times higher for the somewhat more advanced queries.
And also Bing version 2 search API is free!
Whereas, Yahoo BOSS V2 has pricing and so Google does custom search API.
The documentation for Google Documents List API, seems to say that you can create a local document and upload it. Is there no way to actually create and edit a document on Google Docs through an API?
While the docs call it "uploading", everything boils down to sending an appropriately formatted HTTP POST request, so of course it can actually be a new creation rather than an actual "upload" of an otherwise existing file. (Creation through POST requests is similar to what's normally described as a REST API, though in real REST you'd typically use a PUT request instead of course).
You just need to create a blob of data representing your document in any of the formats listed here -- depending on your programming language, simplest may be text/csv for a spreadsheet and application/rtf for a text-document -- then put in in an appropriately formatted POST data. For example, to make a spreadsheet in the simplest way (no metadata), you could POST something like:
POST /feeds/default/private/full HTTP/1.1
Host: docs.google.com
GData-Version: 3.0
Authorization: <your authorization header here>
Content-Length: 81047
Content-Type: text/csv
Slug: Example Spreadsheet
ColumnA, ColumnB
23, 45
Each specific programming language for which a dedicated API is supplied may offer help with this not-so-hard task; for example, in Python, per the docs, the API recommends using ETags to avoid overwriting changes when multiple clients are simultaneously "uploading" (i.e., creating or updating docs). But preparing the POST directly is always possible, since the almost-REST API is documented as the protocol underlying all language-specific APIs.
Alex's answer, while undoubtedly correct, begs the question: "how do I do that via the Google Docs API?"
Here's a way (in Python, 'cause I'm that kind of guy):
import gdata.docs.service
import StringIO
client = gdata.docs.service.DocsService()
client.ClientLogin(username, password,
source='Spreadsheet Creation Example')
content = 'COL_A, COL_B, COL_C, COL_D\ndata1, data2, data3, data4'
ms = gdata.MediaSource(file_handle=StringIO.StringIO(content),
content_type='text/csv',
content_length=len(content))
entry = client.Upload(ms, 'Test Spreadsheet')
This is a small mashup of techniques that I found in http://code.google.com/p/gdata-python-client/source/browse/tests/gdata_tests/docs/service_test.py , which I in turn found via this post from the Google Group for the GData Docs API.
The key insights (for me anyway) were:
realizing that the MediaSource constructor's formal parameter "file_handle" will take any file-like object, and
discovering (as the OP's followup to the Google Group post mentions) that the unit tests are a great source of examples
(I wasn't able to find the Python-specific developer's guide referenced by Alex's doc link -- possibly it's been lost or buried in Google's move of documentation assets from code.google.com to developers.google.com. Alex's link now redirects to the more generic document that shows mostly .NET and Java examples, but only a little Python.)
As of Feb 4, 2019, Google Docs now has a REST API.
See documentation:
https://developers.google.com/docs/api/
(Sep 2019) There are 3 ways to create a document in Google Docs programmatically:
Google Docs REST API (low-level; Python, JS/Node.js, Java, C#/.NET, PHP, Ruby, Go, etc.)
Google Apps Script (high-level; JavaScript-only)
Google Drive API (low-level like Docs API above; both alternatives above can create or edit documents, but this one is create- or delete-only plus editing sharing/permissions)
The Docs API was officially launched in Feb 2019. I produced a high-level video overview of what a mail merge application using the API would look like. (It's not a full-fledged G Suite Dev Show episode but does link to a working sample.) Check out the various guides on using the API, including Quickstart examples in a variety of programming languages.
OTOH, Apps Script is a simpler, higher-level alternative. It's a custom server-side JavaScript runtime supporting apps that are hosted+executed in Google's cloud. Use objects to talk to various Google APIs (G Suite & beyond) without knowledge of HTTP, REST, nor OAuth2. You can also access external databases with its JDBC Service or call other apps via its URL Fetch Service.
With Apps Script, you can create standalone applications, document-bound applications (only works for a single document), or Google Docs Add-ons to extend the functionality of Google Docs. Here are the Google Docs Apps Script overview page as well as the Apps Script reference documentation for Google Docs (Document Service). I've also produced a variety of Apps Script videos if that's your preferred learning vehicle. If you're new to Apps Script, see my answer to a similar SO question for more learning resources.
Typically the Docs, Sheets, Slides, etc., APIs are used to perform document-oriented functionality while the Drive API is used primarily for file-based operations. However "create" is a special case where you can use either. See my answer to another SO question which shows the difference b/w creating a new Google Sheet via the Sheets API vs. the Drive API. (Both samples in Python.) Read this if you're interested in managing sharing or updating permissions of Google Docs.