Add bpipe authentication with an application name #288

alfredkanzler · 2019-01-29T16:14:27Z

Bloomberg's bpipe api allows authentication with an application name as described in "Bloomberg API Version 3.x Developers Guide" (https://data.bloomberglp.com/labs/sites/2/2014/07/blpapi-developers-guide-2.54.pdf) section 6.4.

Added functions to support this form of authentication. Please send questions to [email protected].

with an application name

…bpipe connection with an application name. Description of authenticating a bpipe connection with an app is described in the "Bloomberg API Version 3.x Developers Guide" in section 6.4

johnlaing · 2019-01-30T01:29:32Z

Is there a reason to create separate functions for this case? It seems simpler/preferable to incorporate an optional app name into the existing functions. But perhaps I'm missing something - I'm not familiar with the details of Bpipe.

alfredkanzler · 2019-01-30T15:46:34Z

There were a few reasons that I wrote new functions. Let me first say that I would rate myself as an intermediate level R developer at best. I have not developed full time in over ten years. I developed this because I needed to access bpipe. Now to the reasons

1.) the authenticate C++ function was structurally different enough that I thought it warranted a new function.
2.) I wanted to make my functions as C compatible as possible so did not want to add optional parameters or use function overloading. This precluded naming my blpConnect function blpConnect or adding an optional parameter. This was not a requirement but was my motivation.
3.) The only way I know to overload an R function is via the S3 class mechanism. The authenticate function had required parameters that I did not require to authenticate with an application. I would have had to recognize the desire to authenticate bpipe with an application name and call the correct C/C++ function. I did not know how to recognize that the function should be used to authenticate bpipe.

Those are the structural reasons. I also wanted to isolate my code since I do not consider myself a good package developer. Some of my assumptions above may not have been correct. If these are incorrect please inform me where my knowledge is lacking.

…emoved

…functions were removed.

eddelbuettel · 2019-03-15T18:28:16Z

@alfredkanzler Very cool to this moving forward. I'll take a closer look at the code in the next day or two.

@johnlaing Can you give it a spin to run it at your end?

* release 0.3.9 * more URL updates needed bloomberglags no longer documents API

eddelbuettel

PR looks great -- could not help myself and edited some whitespace, and also added Al to copyright header for those two files

Tried bringing it in line master by cherry-picking our commit, but it still thinks '8 ahead, 1 behind' so squash merge may be easiest.

Builds and checks cleanly on r-release and r-devel so no CRAN issues expected.
Remaining issue: @johnlaing to review and, if possible, run

eddelbuettel · 2019-03-16T12:33:16Z

R/blpConnect.R

+        con <- blpConnect_Impl(host, port, NULL)
+    } else {
+        con <- blpConnect_Impl(host, port, app_name)
+    }


Here the if (is.null(...)) and else branches could be collapsed into one which might be simpler. Yet this is clear.

Or it might be better to change it to two lines including one comment that null/non-null is handled at the C(++) level?

I think this is a good point. I can make the change on Monday. I like the simplification with a note in the documentation.

Sounds good! The code is in great shape but I want to hear from @johnlaing too.

eddelbuettel · 2019-03-16T12:35:19Z

src/authenticate.cpp

+    if(uuid_ == R_NilValue) {
+        identity_p = authenticateWithApp(con_);
+    } 
+    else {


I think we put our else on the same line but hey...

More than happy to follow your preferred format. Since I am going to make the change above I will make the change here.

See the 'Outdated' tag :) While I was at it I pushed some trivial changes back to your branch. So make sure you pull on Monday.

eddelbuettel · 2019-03-16T12:36:43Z

src/authenticate.cpp

+    } 
+    else {
+        identity_p = authenticateWithId(con_, uuid_, ip_address_);
+    }


I like the simpler switch here and the hand-off to these two specialised functions with the old default and new add-on. Nicely done.

johnlaing · 2019-03-18T14:46:16Z

This broke my connectivity right off the bat. I'm going to have to dig in here and figure out what's happening. Bear with me...

alfredkanzler · 2019-03-18T14:51:01Z

Can you send me how you connect? I will also dig into why it failed.

johnlaing · 2019-03-18T17:22:01Z

This was just a SAPI connection, so specifying host/port.

johnlaing · 2019-03-19T09:17:04Z

R/blpConnect.R

 ##' via \code{blpAuthenticate} after \code{blpConnect}.
 ##' @examples
 ##' \dontrun{
 ##'   con <- blpConnect()   # adjust as needed
 ##' }
 blpConnect <- function(host=getOption("blpHost", "localhost"),
                       port=getOption("blpPort", 8194L),
-                       default=TRUE) {
+                       default=TRUE,
+                       app_name = getOption("blpAppName", NULL)) {


Stylistic nitpick: we should use something other than snake_case for argument names. We've tended toward camelCase in most functions, thought there are also a few old-school R dot.separated names left over. I think appName fits best here.

johnlaing · 2019-03-19T11:02:33Z

src/blpConnect.cpp

+// [[Rcpp::export]]
+SEXP blpConnect_Impl(const std::string host, const int port, SEXP app_name_) {
+    Session* sp = NULL;
+    if (app_name_ == NULL) {


I think this should be R_NilValue rather than NULL. That seems to be what was killing my connection yesterday.

alfredkanzler · 2019-03-19T12:50:29Z

I missed fixing this. I will make the change and change to camel case. I should have this done today.

updated documentation and removed redundant code

clarification of appName usage.

Previously tested this with == NULL.

alfredkanzler · 2019-03-19T21:09:38Z

I changed to camel case in the R code. I also think the bug is fixed. John should test.

johnlaing · 2019-03-20T17:37:56Z

Bug is fixed! This is great, thanks for your work. Just a couple other cleanup items from me:

I don't think we need separate functions for blpConnect with/without app. The NoApp version is a subset of the WithApp, I'd be in favor of combining. (This is in contrast to authorization, where the internals are completely different and separate functions make perfect sense.)
In the R blpAuthenticate function we should pull the IP conversion inside the case statement, as it's irrelevant for the BPIPE code path.

I can take care of these easily on my end. And if we're all in agreement I can push back to your branch, Al, for one more quick test before merging.

alfredkanzler · 2019-03-21T09:40:59Z

This is good news. I unfortunately cannot test today. I will try to get setup to test tomorrow.

eddelbuettel · 2019-03-22T10:48:24Z

I made some teeny edits to ChangeLog (and DESCRIPTION) and resolved the merge conflict. Let's do the one final test and then it should be good to go to the repo ... and CRAN!

eddelbuettel · 2019-03-22T11:08:01Z

Works for me. No need to rush this now that we did it so well.

alfredkanzler · 2019-03-26T13:00:57Z

I just completed testing without any issues in my tests.

eddelbuettel · 2019-03-26T13:30:05Z

I just merged in a dry-run in another repo and the graph is pretty impressive so I may change my plan and not not squash-merge. Any objections?

alfredkanzler · 2019-03-26T13:33:58Z

No objections from me

eddelbuettel · 2019-03-26T13:59:09Z

@johnlaing All green from your end too?

johnlaing · 2019-03-26T14:25:58Z

All good here, in the sense that the code works. My preference is to always maintain commit history unless there's a very good reason to squash.

eddelbuettel · 2019-03-26T14:37:41Z

All good here, in the sense that the code works.

Ace. Will merge and release to CRAN (this eve or tomorrow eve). I'll probably call it 0.4.0 for good measure.

My preference is to always maintain commit history unless there's a very good reason to squash.

Yes, understood. It is a trade-off. In a squash the log of the individual commits become part of the one commit's log. So not erased, just a wee bit less visible. The upside is the saner graph. With a lot of concurrent activity those become harder to use.

alfredkanzler added 2 commits January 28, 2019 16:11

exposed R function that allow users to authenticate a bpipe application

7d7ea83

with an application name

This have been added or modified to allow the user to authenticate a …

a49b277

…bpipe connection with an application name. Description of authenticating a bpipe connection with an app is described in the "Bloomberg API Version 3.x Developers Guide" in section 6.4

alfredkanzler added 3 commits February 12, 2019 09:37

fixed some documentation

ef15196

bpipe access was added to the blpConnect method. The new method was r…

a0ac289

…emoved

bpipe access was added to blpConnect and blpAuthenticate. New access …

55298bb

…functions were removed.

eddelbuettel added 3 commits March 16, 2019 08:04

release 0.3.9 (closes Rblp#290) (Rblp#291)

4d81364

* release 0.3.9 * more URL updates needed bloomberglags no longer documents API

minor whitespace edits, roll minor version

d1d5e4a

revert roll minor version

4c6289f

eddelbuettel approved these changes Mar 16, 2019

View reviewed changes

johnlaing reviewed Mar 19, 2019

View reviewed changes

alfredkanzler added 4 commits March 19, 2019 15:53

changed naming case to match original authors conventions.

97b7942

updated documentation and removed redundant code

comment changes to reflect change in naming conventions and some

8b38f80

clarification of appName usage.

updated version number to 3.9.2

6350cc2

fixed bug where R based NULL is tested in C++ with == R_NilValue.

d86f4fe

Previously tested this with == NULL.

documented changes

b77f89c

johnlaing and others added 4 commits March 21, 2019 07:38

only do IP resolution on SAPI path

1f29421

streamline connection code

42c656e

edits to ChangeLog

f932bd1

Merge branch 'master' into master

9fd54f6

eddelbuettel merged commit 9b75403 into Rblp:master Mar 26, 2019

rharlow86 mentioned this pull request Jun 24, 2021

Extending B-PIPE Authentication Methods #342

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add bpipe authentication with an application name #288

Add bpipe authentication with an application name #288

alfredkanzler commented Jan 29, 2019

johnlaing commented Jan 30, 2019

alfredkanzler commented Jan 30, 2019 •

edited

Loading

eddelbuettel commented Mar 15, 2019

eddelbuettel left a comment

eddelbuettel Mar 16, 2019

alfredkanzler Mar 16, 2019

eddelbuettel Mar 16, 2019

eddelbuettel Mar 16, 2019

alfredkanzler Mar 16, 2019

eddelbuettel Mar 16, 2019

eddelbuettel Mar 16, 2019

johnlaing commented Mar 18, 2019

alfredkanzler commented Mar 18, 2019

johnlaing commented Mar 18, 2019

johnlaing Mar 19, 2019

johnlaing Mar 19, 2019

alfredkanzler commented Mar 19, 2019

alfredkanzler commented Mar 19, 2019

johnlaing commented Mar 20, 2019

alfredkanzler commented Mar 21, 2019

eddelbuettel commented Mar 22, 2019

eddelbuettel commented Mar 22, 2019

alfredkanzler commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019 •

edited

Loading

alfredkanzler commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019

johnlaing commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019

Add bpipe authentication with an application name #288

Add bpipe authentication with an application name #288

Conversation

alfredkanzler commented Jan 29, 2019

johnlaing commented Jan 30, 2019

alfredkanzler commented Jan 30, 2019 • edited Loading

eddelbuettel commented Mar 15, 2019

eddelbuettel left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

johnlaing commented Mar 18, 2019

alfredkanzler commented Mar 18, 2019

johnlaing commented Mar 18, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alfredkanzler commented Mar 19, 2019

alfredkanzler commented Mar 19, 2019

johnlaing commented Mar 20, 2019

alfredkanzler commented Mar 21, 2019

eddelbuettel commented Mar 22, 2019

eddelbuettel commented Mar 22, 2019

alfredkanzler commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019 • edited Loading

alfredkanzler commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019

johnlaing commented Mar 26, 2019

eddelbuettel commented Mar 26, 2019

alfredkanzler commented Jan 30, 2019 •

edited

Loading

eddelbuettel commented Mar 26, 2019 •

edited

Loading