Using Azure Logic Apps to get started with our tool

By: Tyler Simpson | Date: 07/23/2024

Introduction

At Granum Technologies LLC, our mission is to modernize the fintech industry by making powerful software tools accessible to as many organizations as possible. Our API creditparsepro.io aims to do this by allowing you to parse fixed-length credit reports into actionable curated fields ready for analytics and machine learning.

Traditionally, interacting with APIs required engineering expertise, but this is no longer the case. In this demo, we will show you how to leverage Azure Logic Apps to interact with our API without writing a single line of code.

While this guide offers a comprehensive approach to setting up the necessary Azure services, it does not cover networking and security best practices, which are critical for enterprise-level solutions. The logic demonstrated here using Azure Logic Apps can also be replicated in most RPA tools, including Microsoft Power Automate, which is included with Microsoft 365 licenses and is very similar to Logic Apps.

Services

If you already have an existing Azure architecture, you can skip this section. Here, we'll outline how to create an Azure environment and all the services required for this demo. Resource and service names will be italicized, specific names given to them will be underlined, and page section names will be in 'apostrophes'.

Because of my location, I default to using the 'East US' region for my services. You should consult your closest data center here.

Please note that when naming my Azure resources I follow the Azure recommended best practices here.

Services covered:

Azure Account
Azure Billing Account
Azure Subscription
Azure Resource Group
Azure Key Vault
Azure Data Lake Storage Gen 2
Azure Logic App

1. Create an Azure account

If you do not have an Azure account, create a free account here.

Azure offers a generous $200 credit for first-time users.

During account creation, you will also set up a Billing Account.

Note that even if you do not receive the $200 free credits this demo will cost only a few pennies. If you are still in the 1st 12 month period of a new account the free resources included in that will likely reduce the cost of this demo to 0.

2. Create a Subscription

A Subscription is basically a logical grouping of services, you can name it and attach it to your Billing Account.

Pricing: This is a free resource.
Docs

3. Create a Resource Group

A Resource Group is another logical grouping of services, I tend to use it for environments i.e. test, prod.

Pricing: This is a free resource.
Docs
rg-grtech-test-001.
- rg - Resource Group
- grtech - Granum Technologies
- test - Environment (dev, test, prod, etc)
- 001 - Increment for future version improvements

4. Create a Key Vault

A Key Vault is used to securely store keys and secrets, we will be using it to securely store our API key.

Pricing: 12 months free with a new account other $0.03 for 10,000 transactions.
Docs
Upon signing up on Stripe you would have received an API key via Email.
kv-grtech-test-001
Once created, create a secret and add your API key as the value.
cpp-x-api-key

4. Create an Azure Data Lake Storage V2 (ADLS2)

ADLS2 is a "blob" or object storage service, similar to a local file system but in the cloud. We will use it to store both credit reports for parsing and the parsed reports.

Pricing: 1 year 5gb storage free with new accounts. Pricing is very low and competitive, read more here.
Docs

There are 2 important settings that you need to check when creating your ADLS2.

1st you want to ensure you 'Enable hierarchical namespace' on the 'Advanced' tab. This is what makes the storage account a blob storage service. Changing this after creation can be difficult.
2nd you likely want to ensure in the 'Networking' tab that you DO NOT 'Enable public access from all networks'. You are working with very sensitive information. In my case, I am using example reports that are not associated with real people but I am still taking precautions. For this example you can use 'Enable public access from selected virtual networks and IP addresses' as a rudimentary form of security. You should be in contact with your networking and security team if you have one. You can change this setting after creation.
- Your client IP address should be listed in the 'IP Addresses' section allowing you to white-list it.
stcppgrtechtest001
Now that the storage account is created we want to create folders/objects for our files
'Storage Explorer' > 'Blob Containers' > 'Add Container'
- applications/072024/
- applications-cleaned/072024/

Note that in this demo we are replicating credit reports as coming in on JSON files. In this example multiple credit reports can come over in a single file and each have an associated applicationNumber. This is likely much simpler than how you receive them and only illustrates one simple possibility.

[

{

"applicationNumber": "101",

"report": "report text here"

},

{

"applicationNumber": "102",

"report": "report text here"

}

]

5. Create a Logic App

A Logic App is primarily a no-code tool that allows users to create logical operation flows. It has many built in connectors and a simple interface.

Pricing: https://azure.microsoft.com/en-us/pricing/details/logic-apps/. You will get 4,000 actions free and after that they are only $0.000025 per action.
Docs
Choose the consumption option since we are not creating an enterprise solution and this will be the far cheaper option costing nearly nothing to run.
la-cpp-grtech-test-001

6. Settings

In the Logic App settings, navigate to 'Settings' > 'Identity' > 'System assigned' and turn it on.
- This is because we will be using this identity to allow the Logic App to securely access the other Azure services.
At either the Resource Group or Subscription level you will now want to set IAM permissions for this identity we created. Permissions are inherited by default so if you set a permission at the Subscription level it will be inherited by all Resource Groups inside it.
- Navigate to your Resource Group then to 'Access control (IAM)' > '+ Add' > 'Add role assignment'. We will be adding 2 roles to our Logic App identity.
  - Role: 'Key Vault Secrets User'
  - Members: 'Managed Identity' select the Logic App
  - Role: 'Storage Blob Data Contributor'
  - Members: 'Managed Identity' select the Logic App

The Logic App now has access to the Key Vault and the ADLS2 Storage Account.

Logic App

Now that the Azure services are set up and the correct permissions have been granted, we can begin building our Logic App. You can edit via the GUI or code view. I will provide images and examples as we proceed focusing on the GUI editor.

Overview

1. Trigger the workflow when a JSON file containing credit reports enters our ADLS2 blob storage.

2. Retrieve the API key secret from our Key Vault.

3. Get the JSON contents of the created blob.

4. For each record in the JSON blob, send it to creditparsepro.io API and get the return values.

5. Write the return values to a new blob in our ADLS2 blob storage.

1. Trigger

Add 'When a blob is added or modified'.
You will be prompted to create a connection. For 'Authentication Type' choose 'Logic Apps Managed Identity'. I set the name to the name of the storage account stcppgrtechtest001.
Add the ADLS2 endpoint which can be found in the ADLS2 'Settings' > 'Endpoints' > 'Data Lake Storage'.
Add the Container path.

2. Key Vault

Add 'Get secret'.
You will be prompted again to create a connection. Again choose 'Authentication Type' of 'Managed Identity' and add the 'Vault Name' created earlier kv-grtech-test-001.
Add the name of the secret created earlier cpp-x-api-key.

3. JSON contents

This step will be relatively complicated if you are new to Azure and Logic Apps/RPA. When we get the Blob content you may expect it to be in JSON format but it is actually Base64 encoded. There is currently no built in action so we will need to use a 'Compose' action which will allow us to customize how we manipulate the data. Azure allows a built in function that we can call from within the 'Compose' activity to convert this Base64 into a string (text). Then once we have the JSON string we need to parse it by showing the Logic App the correct file structure.

3.1 Get blob content

Add 'Get blob content' activity.
Add ADLS2 endpoint as we have done before.
Add blob name dynamically to the end of the file path.
- In 'Blob' when you click you will see a lightning bolt button.
- Select the 'List of Files Name' output from the trigger 'When a blob is added or modified'.

Now we have the Base64 encoded data from the blob picked up by the trigger.

3.2 Base64 to String

Add 'Compose' activity.
In the input click the 'fx' button for custom input.
base64ToString(body('Get_blob_content')['$content'])

Now we have the JSON data in string format.

3.3 Parse JSON

Add 'Parse JSON' activity.
Add blob name dynamically to the end of the file path.
- In 'Blob' when you click you will see a lightning bolt button.
- Select the 'List of Files Name' output from the trigger 'When a blob is added or modified'

Now we have the Base64 encoded data from the blob picked up by the trigger.

You can pass the example JSON schema I gave earlier or use your own file schema.

[

{

"applicationNumber": "101",

"report": "report text here"

},

{

"applicationNumber": "102",

"report": "report text here"

}

]

4. For each

Remember that in our scenario there can be multiple credit reports or "applications" coming in on each JSON file. Because of this we need to loop through each record that was parsed out of the JSON file and handle them individually.

Add 'For each' loop.
In 'Select An Output From Previous Steps' add the 'Parse JSON' 'Body'.

5. HTTP (API)

Add 'HTTP' within the 'For each' loop.
Add the appropriate Headers and Body format based on the creditparsepro.io docs.
For the 'x-api-key' Authentication Header add the 'Get secret' 'value'.
For the body add the 'Parse JSON' 'Body report' to grab the report.

6. Create blob

Add 'Create blob' within the 'For each' loop.
Add the ADLS2 endpoint and folder path.
Dynamically set the blob name as the applicationNumber from 'Parse JSON'
Dynamically set the blob content with the 'HTTP' (API) body.

As we wrap up our Logic App build I want to mention that this is a basic framework that can and should be expanded upon.

If you look at the creditparsepro.io postman docs you will notice we offer GET operations for both your current rate limit and monthly quota limit. A more robust solution could include adding actions to GET rate and quota limits with logic gates to avoid unexpectedly reaching a quota.

Further, adding error logging, handling, and notifications would be a good idea. You could use Azure table storage in your storage account to log errors and when they occur as well as send Emails to key stakeholders.

Input Output

Now that our services are created and we have built our Logic App let's take a look at the input file(s) and output file(s).

[

{

"applicationNumber": "101",

"report": "1100028022924144736TTXA0700@13000230670P10113033AG@3220018*666048258@335004216/JOE/M/EARNEST//1928E61208161928@335004324/JOSEPH/MARCEL/EARNEST// E105N@3360078080609232163182310S 41/1101/CORRAL/ST///SWEETWATER/TX/795566611@336007304230423600 36/CORRAL COVE/////SWEETWATER/TX/79556@337008608130813223SIEMENS STROMBERG CARLS15400 RINEHART RD12LAKE MARY FL01 32746 @357018931N06221023 07REV100000600L00000023H100323 11 CR0001 A913117600189B2141600000000B420B00000000000000CB520 071222B6261389998DC11THE BON TONF316102305110740@3570185 P01210223 26360100067400O 020723 05 CI0001 A9118005758B2142300000000B427BCCCCCCCCCCCCCCCCCCCCCCB6372991785FM22PRINCIPAL RESIDENTL MTF3160223051126 @357016619N04971022 15REV100003000L00002048H101122 12 CR0001 A9082830B2149900000000B429B000000000000000000000000B6191102240BB04BANKF3161022A2111519@357018418N04021221 18REV100003200L 121621 12 CR0001 A916542418054042B2149900000000B429B-------------00-000-----B6291240000BC14CITICARDS CBNAF3161221A2111818@357018419N09051221 18REV100000000L 121521 12 CR0001 A916542418053826B2149900000000B429B--------------000-----00B6291240000BC14CITICARDS CBNAF3161221A2111819@357018419N10040621 18REV100005400L00004885H061321 12 CR0001 A916545715001233B2148500000000B429B000000CC00000000CCCC000CB6293278165BC14CITICARDS CBNAF3160621A2111819@357016519N08010920 18REV100000800L00000025H091120 12 CR0001 A9147791556597B2141200000000B416B000000000C0B6251232920BC10SHELL/CBNAF3160920A2111819@357018319N05040520 18REV200007500L00004569H051320 12 CR0001 A916601100462050B2142900000000B429B000000000CCCCCCCCCCCCCCCB6283276502BC13DISCOVER BANKF3160520A2111819@357016519N04970320 15REV100003000L00002048H031120 12 CR0001 A907383B2142500000000B429BCCCCCCCCCCCCCCCCCCCCCCCCB6191102240BB04BANKF3160320A2111519@357018118N11950519 18REV100002831H 051319 12 CR0001 A916529107125292B2149900000000B429BCCCCCCCCCCCCCCCCCCCCCCCCB6261270246BC11CAPITAL ONEF3160519A2111818@3570180 P11130515 27014200032000O 052815 12 CI0001 A9149590103426B2142000000000B423BCCCCCCCC---CCCCCCCB6332184720BB18HOME LOAN SERVICESF3160515A21127 @3570197 P05931223 07REV100000200L00000959H1225230000000011 OR0001 A9141687725307B2149900000000B4290000000000000000000000000B520 121714B6243321860DC09SYNCB/JCPF3161223A11107 @3570204 P05011223 18REV200010200L00010389H1222230001038911 OR0001 A9204118160281358112B2147400000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000273S 110923B6253182310BC10JPMCB CARDF3161223A11118 @3570191 P09221223 07REV100002150L00000037H1217230000000011 OR0001 A916504994014753B2141600000000B420000000000000000CB520 092822B6251323180DC10SEARS/CBNAF3161223A11107 @3570197 P11861223 18REV200005200L00013887H1216230000539811 OR0001 A9087264B2146600000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000171S 110523B6301230206BC15BANK OF AMERICAF3161223A11118 @3570205 P07951223 18REV100008500L00008398H1214230000838311 OR0001 A916433600615056B2144800000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000132S 120623B6303202754BC15BANK OF AMERICAF3161223A11118 @3570197 P05011223 18REV200011400L00011330H1211230001131311 OR0001 A9084394B2149000000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000343S 120623B6301213727BC15BANK OF AMERICAF3161223A11118 @3570204 P05031223 18REV100008000L00007635H1210230000746011 OR0001 A9205323510002673205B2149900000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000149S 120623B6251240338BC10JPMCB CARDF3161223A11118 @3570202 P11041223 18REV100005000L00005029H1209230000490211 OR0001 A919324131017164644B2149900000000B429CCCCCCCCCCCCCCCCCCCCCC000B52000000132S 120523B6240203010BC09FNB OMAHAF3161223A11118 @3570204 P11911223 18REV100015000L00014669H1205230001111811 OR0001 A9205121071814310177B2149900000000B429CCCCCCCCCCCCCCCCCCCCCCCCCB52000000232S 120223B6251230750BC10SEARS/CBSDF3161223A11118 @3570187 P12201223 26360100067400O 1205230005491011 OI0001 A9118005758B2141100000000B415CCCCCCCCCCCB52000000595S B6312570635FM16CITIMORTGAGE INCF3161223A11126 @3570181 P07231223 18REV100001000L00000018H1203230000000011 OR0001 A916042378879907B2140500000000B4090CCCCB520 081123B6262310220DC11KOHLS/CHASEF3161223A11118 @3570199 P12201123 18REV100003500L00003509H1128230000349711 OR0001 A9140007161206B2143600000000B429CCCCCCCC-0000000000000000B52000000121S 110423B6263279024BC11CAPITAL ONEF3161123A11118 @3570188 P04211221 07REV100000118L00000118H1201210000000011 OR0001 A913211478102B2140800000000B4120000000CB520 050721B6331362760ZR18BOSCOVS DEPT STOREF3161221A11107 @3590050081123 UNKNOWN31UNKB6231219240BC08INFIBANK@3590053070823 UNKNOWN31UNKB6261352920DC11KOHLS/CHASE@3590062102722 UNKNOWN31UNKB6351252660AU20700 CREDIT/AA MOTORS@3590063090422 UNKNOWN31UNKB6361352085DC21COMPLETE DEPT. STORES@361005457420335 AG08TOO MANY INQUIRIES LAST 12 MONTHS@361003757251202 OFAC NO RECORD FOUND@361003657241204 MLA NO RECORD FOUND@8500101000000N/A 00054910000624600000000000001553 00000595 011 004001024011024000000118600000000@950001604105385@"

},

{

"applicationNumber": "102",

"report": "TU4R062011 1614F 0459317220240229143215PH0101207000SH01027101Y05N03PEN20210602NM01070F1DAY SUNNY PI01029F66688567819910601 AD01105F11333 HOPP CT FANTASY ISLAND IL60750 20221206AD01105F111233 S CARROT ST FANTASY ISLAND IL60750 20200408EM01100ASDF F 20240110V EM01100DEFI SLUTIONS FDEVELOPER 20240110V TR01288Q 0202V002WLM P & F CU I44009 I2020042620240101V 051000003817000006587 034M000000242 US 00 2024010105 00 TR01288B 041PF002FST PREMIER R38294 I2022060120240101V 011000000184000000414000007500 00 20231201111111111X11 120000002TR01288B 09616003DISCOVERBANK R4 I2012042520240101V 011000000093000000093000008000MIN 000000010 CC 00 2023120111111111111111111111111111111 290000002TR01288Q 06262001PROVDENT FCU I75500045 I2020100420240101V 011000000554000001005 00 202312011111111111111111111111111 250000002TR01288B 0848R015WILM TRUST M698007 I1995013120240101V 011000003327000036500 300M000000163 RE 00 20231201111111111111111111111111 240000002TR01288D 0160Q001BOSCOVS R1 I2007121020231201V 011000000019000000272 MIN 000000010 00 202311011XXXXX111111XX1XXX111X1XXXXXXXXXXXX111111XX1X11 470000002TM01040000000000001000000000005000000000IN0106516ORF 04593172NEEDHAM/RECF I 20240229IN0106540PEY 02462756MIDLAND CRED I 20240113OB01148TRANSUNION TEST FACILITY 555 W ADAMS CHICAGO, IL 60661 800-888-4213 AO0101700Q8804 SC0103400Q88+696 039013027018 AO010170680004O01AO010170705101M00ENDS010024"

}

]

Output

We are looking at just 101.json since 102.json will be somewhat redundant.

{

"address": {

"address_changes_per_year": 0.0,

"change_count": 0,

"months_at_current_address": 210.96,

"state": "tx",

"zip": 79556

},

"age": 95,

"collections": {

"count": 0

},

"date": "2024-02-29",

"delinquencies": {

"past_due": {

"open_count": 0,

"open_sum": 0,

"total_count": 0

},

"times_late": {

"30_days_sum": 0,

"60_days_sum": 0,

"90_days_sum": 0

}

},

"dti": 0,

"employment": {

"months_at_current_employer": 126

},

"inquiries": {

"past_6_months_count": 1,

"total_count": 4

},

"public_records": {

"bankruptcy_count": 0,

"count": 0

},

"score": 670,

"tradelines": {

"credit": {

"open_count": 3,

"total_count": 4

},

"installment": {

"installment_utilization_ratio": 0,

"open_count": 0,

"open_sum": 0,

"total_count": 0

},

"mortgage": {

"open_count": 1,

"total_count": 3

},

"overall": {

"oldest_open_date_months": 447,

"open_accounts": 13,

"total_accounts": 24,

"total_monthly_payments": 1553

},

"revolving": {

"open_count": 9,

"open_sum": 67800,

"revolving_utilization_ratio": 92.12,

"total_count": 17

}

Conclusion

In this demo, we have demonstrated how easy it is to interact with our creditparsepro.io API using Azure Logic Apps, a no-code tool. This approach eliminates the need for engineering expertise, making powerful fintech tools accessible to a wider audience. While this guide covers the basics, remember to consider networking and security best practices for enterprise solutions. The same logic can be applied using other RPA tools like Microsoft Power Automate, providing flexibility and ease of use for organizations of all sizes.