Transform Your Workflow with Our Invoice Model PDF for Quality Assurance

Invoice model pdf for Quality Assurance

hi guys welcome back to the session today we will create a uipath project to extract data from different invoices that means this project should work even if you have invoices of different formats and for this project we will use a combination of machine learning model present in AI Fabric and document understanding package so let's get started the first step of this project would be to create ml skill under AI fabric of uipath in case if you are not aware of what is AI Fabric and how to set up AI Fabric in uipath please check out my introduction video on AI fabric so here I am inside uipath automation Cloud let's move to AI Fabric and here we will create a new project so for that click on plus sign let's give a project name here invoice data extraction let's give a description here this project will extract data from different invoices and let's create it now here first of all we'll create an ml package so for that move to ml packages and to use an already existing machine learning model we will move to this out of the box packages so let's open this one and here we have multiple of options present for our case we'll go with this uip part document understanding here also we have different machine learning models present for our case we will go with this invoices machine learning model this invoices machine learning model is used for extracting commonly occurring data points from invoices including header fields and line items now what are common only occurring data points in invoices so let me show you an example here this is a sample invoice with me now here if you see the common data points in any invoice would be the invoice number invoice date the company name the due amount and so on so we will be extracting these data points from different invoices that is the invoices of different format let's move to ml packages again this is the invoices ml package on which we will work so for that choose a package version from here we have multiple of package versions present so let's go with the latest one that is 4.0 and let's hit submit now here will'll provide some information package name let's provide package name as package invoices we have chosen the package version as 4.0 this is the description in input description output description now here we can select an OCR engine so from the drop down from this drop down we can select any OCR engine I'll go with this uip paath OCR next option is to provide OCR URL let's skip this one this is optional and next is the option to provide OCR API key now to get this OCR API key what you can do you'll move to this admin licenses and under licenses let's move to other services and here under document understanding we have this API key we have this API key so let's copy this API key from here and I'm going to paste it here and let's hit submit so we created this ml package that is p invoices our next step is to create the ml skill so let's move to ml skills and create new skill here let's give a ml skill name so I'll write ml invoices let's choose the ml package from here that is p invoices which we just created next let's choose package major version which is four let's choose package minor version which is zero we can provide a skill description here so let's provide the description as ml skill to extract invoice data and since I'm am using interprise trial Edition I do not have gpus so I'm not going to enable this GPU all details has been provided let's create this ml skill so this ml skill has been created the status you can see it is still deploying it will be deployed to the orchestrator so let's move to orchestrator to check this one so here let's move to home and let's open the orchestrator and let's move to tenant now under this tenant we have the option of ml skills now here you see this ml invoices skill which we have just deployed here the status is deploying after some time the status will be changed to available once the de once the deployment is complete so we are done with the first part we have created this ml skill using the invoices machine learning package now our next step of the project would be to create the project in UI Parts Studio which will extract the invoices data with from the invoices of different format so let's open this main workflow and here I will use the workflow type as flowchart so let me add this flowchart activity here let's open this flowchart and here I have a folder which contains invoices of different formats so let me show you the samples this is an invoice of one format and this is an invoice of different format so we wish to extract data from invoices of different formats the data which will be which we will extract our invoice number due date balance due and this company name and since we wish to work on all these invoices first of all I'll store the file path of all these invoices inside an array variable so let's move to UI path and here I'm going to use this assign activity let me create a variable also so first variable I'm going to create is folder path and I'm going to provide this folder path let's provide it as a default value and in this assign activity under value I'm going to provide the expression as directory dot get files directory. getet files and inside the brackets I'm going to provide this folder path variable which we just created so this expression is going to fetch all the file Parts present inside this folder path variable inside a variable here so I'll create a variable say file paths and this will be a variable of type array array of type string so this is done this file path variable contains the file Paths of all the files present inside this folder now our next step here is to add a sequence and inside this sequence we'll add this for each activity to Loop through all the file paths present inside this variable so here open the sequence and add this for each activity and here I will write for each file present inside the file paths variable so for each file present inside this file path variable we want to extract the invoice data now here to extract the invoice data our first step is to a package called intelligent OCR so let's move to manage packages let's move to official and search for intelligent OCR so this is the package uip part. intelligent OCR do activities this package provides the infrastructure to enable document processing in it allows you to load taxonomy digitize documents classify documents and many other activities so let's select this one and this package and let's hit save so the package has been installed and here you notice as soon as the package is installed this taxonomy manager appears so let's launch this taxonomy manager so this this is the taxonomy manager where user creates document type under a group and category so let's create a group here let's give the group name as G invoices let's create this one next let's create a category so let's give the category name as category invoices and let's create this one also now we'll create a new document type under this group and this category so let's click on add new document type let's keep the document name as stock invoices this is the group and this is the category which we created now here we'll add the fields which we wish to extract from the invoices so click on this new field let's keep the name here so I want to extract the invoice number next we will select the type of the invoice number now here we have different options here I will go with this text type because the invoice number is not necessarily a number it could be in the format of imv 123 and so on so I'm going to select the invoice number type as text and let's hit save let's add a new field and here I will give the name as due date and the type will be date here let's hit save let's add a new field and this time let's give the name as total amount we want to extract the total amount from the invoice as well the type will be number here let's hit save let's add the last field as company which will have the type as text and let's hit save so we created this document type of name doc invoices of this group and this category and all these fields that are invoice number due date total amount and Company let's hit save so we are done with our work under the taxonomy manager let's close this one and now if you move to uip paath project folder here you see inside this project folder this document processing folder got created where we have this taxonomy file so whatever details we entered inside the taxonomy manager just now for that the taxonomy manager created this Json file so this is the Json file which got created by the taxonomy manager of all these details now our next step is to convert this Json file into a variable so for that I'm going to use an activity here that is load taxonomy so let me search for the load taxonomy activity here so this is the activity under intelligent OCR package this activity turns the taxonomy do Json file which we just saw this activity converts this Json file into a variable so let me add this activity here and the Json file which will be converted into a variable will be stored inside an output variable so create a variable here let's keep the name as taxonomy now here you see inside the variable this taxonomy variable is created of variable type document taxonomy now our next step is to digitize the document for which I'm going to use the activity digitize document again this is an activity under this intelligent OCR package let me add this digitize document activity here this activity digitizes a document by extracting its storm and text and store them inside their corresponding variable types so here we'll have to provide the document path which for want to digitize so here I'll provide the document path as file this file contains all the file paths one by one as the loop iterates so I'll provide here the document path as file and here we'll have to convert this into string as well so put a dot and select two string now since this activity digitizes the document by extracting his document text and document object model so we get these two as a variable so create these variables here let's say the variable as Dom and this also let's create this variable let's keep the name as doc text so now the variables has been created and here we have to drop an OCR engine to perform this activity so here I'm going to use the Omni page OCR engine for which we have to another package so move to manage packages and here I'm going to search for Omni page so so this package is present uipath omnip page. activities so I'm going to this and let's hit save so the package has been installed let me search for the Omni page OCR engine here so this is the Omni page OCR let me add this here so all details has been provided inside this digitize document activity let me collapse this our next activity here is to extract the data for which we are going to use the activity data extraction scope so this is the data extraction scope activity which provides a scope for extractor activities so let me add this activity here and here we'll have to provide some details document path and we'll have to convert it into string so put a dot and select two string here next we'll have to provide the taxonomy variable here so let me provide the taxonomy variable next here I will provide the doc text variable which we created inside this digitized document activity next we'll provide the document object model as Dom next we'll have to provide this document type ID this document type ID you will get from this taxonomy Json file which I showed you a few minutes back so here you see this is the document type ID this is the complete ID let's take this from here so let's provide it inside the double codes and this data extraction scope provides the output of extraction results so create this extraction results variable extraction results next here we'll have to drop the extractors here in this project I'm going to use the machine learning extractor for which we'll have to again a package so for that move to packages and here I'm going to search for document understanding package so this is the package uip part. document understanding. ml. activities I'll this one and hit save now as we this package we will get the machine learning extractor so let me search for the same machine learning extractor is present under the extractors so let me add this here so these are the terms and conditions I agree and here we'll have to provide the ml skill the ml skill which we created in the very beginning of the session so if you move to the orchestrator again first of all this is the ml skill which we created ml invoices it must be deployed to the orchestrator so you see under this ml skills this ml invoices has been present and the status has been changed to available that means we can use this ml ml skill now that is ml invoices so for that move to uip path again and here from the drop down let's select this ml invoices ml skill and again here we have to provide the API key which we used earlier as well for that also what we will do we'll move to admin then licenses other services and here we have the API key let's copy the API key and provide it under this API key field and let's hit get capabilities next we'll have to configure the extractors we have added this machine learning extractor with all the details next we'll have to configure the extractor so let's open this one let's expand this and here we will do the mapping of document types and Fields to this machine learning extractor so if you expand this here you see all these fields we have added inside the taxonomy manager under this stock invoices so here we'll do the mapping so for the invoice number the mapping will be this invoice number for due date the mapping will be this date for total amount the mapping will be this total and for company we can simply choose this name so the mapping has been done let's hit save and we are done with this data extraction scope as well so here there is an error we have not provided this Dom so let me provide this variable Dom here and all done under this data extraction scope let's collapse this one so we added this activity for data extraction next to review and correct the data extraction wherever required we will add the activity present validation station so let me search for this present validation St station here so this is the present validation station activity it will open the validation station which enables the users to review and correct the document extraction or document classification so let me add this present validation station activity here here also we'll have to provide all these variables which we already created earlier so let me provide the document path document path will be file dot two string document text here will be Doc text variable document object model will be Dom variable then we have the this taxonomy as well next we'll have to provide the automatic extraction results so this will be the extraction results variable so this activity takes all these input and provides us the output as validated extraction results so I'm going to create this output variable validated results so from this activity we get this validated results variable which we want to convert into data set variable for further use so I'm going to use the activity here export extraction results so let me add this export extraction results activity here here it takes the extraction results which is nothing but the validated results variable and in return it provides the variable into a data set variable so create a variable say data set so now we have complete data extracted from different invoices inside this data set variable now let's say that I want to store the extracted data from this different invoices inside an Excel sheet so how to do that first of all I'm going to convert this data set variable into a data table variable which can be written to the Excel file so let me collapse this export extraction results and here I'm going to add for each activity let me add this one here and here I will provide for each item in for each item in data set do tables so this expression data set. tabls return a data table variable so this item is a data table and ingly the type argument I'm going to change here to system. dat. datat table so for each item in data set. tabls what we want to do do we want to write into the Excel file so for that I'm going to use the append range append range activity from the work workbook package I'm not going to use right range activity since that is going to override the already existing data that's why I'm going to use this append range from the workbook package and here will provide the workbook path so I'm going to provide the workbook path of this invoice data file where we want to store the extracted data from different invoices so browse and from this project folder I'll select this invoice data file and here we can provide the sheet name we can provide the sheet name in the name of the data table name so here I'm going to provide item dot table name item. table name and let's convert it into string so put a dot and select two string so the sheet name will be as item. name. two string and the data table what will be the data table so this item is itself a data table which we want to append so I'll write here item so this for each is also done for each item in this data set do tabls it's going to append the data table inside the sheet of the Excel file now to check what values we have inside the data table you can also add the output data table and message box to display whatever values are present inside the data table variable so our project is ready let's move to flowchart let's save this project and I'll quickly run this one to see how it works now before that let me close the already open invoices we do not require to open the invoices and let's close this file as well and let's quickly run this project and my project is running now and here you see uip paath open the validation station for the first invoice for us to review the extracted data so here you see the data has been extracted invoice number correctly extracted due date is also correctly extracted let me match the due date so this is the due date here due date is also extracted correctly then this is the total amount which is is present here total amount is also extracted correctly and the company name is also correct so let's check this one and let's save this next it will open the validation station for the second invoice so here we got the validation station for the invoice 2 this is the invoice number again which is correct due date is also correct total amount is also correct and the company name is correct let's check this one and let's save this next we will get the validation station for the third invoice so here is the validation station for the third invoice and here you see even if the invoices are of different formats the data is extracted correctly this is the invoice number which is correct this is the due date which is also correct this is the total amount so this also looks good and the company name in this invoice is again a company name so this also looks good so this is the correct extracted data from for the invoices let's save this one and we'll get the last validation station for the fourth invoice and again we got the extracted data from another invoice of different format this is the invoice number which is extracted correctly this is the due date June 16 2019 which is extracted in this format this is also correct the total amount let me check here the total amount also looks good the company name it extracted two times as the company names are present here two times so here I'm going to remove this company name which was repeated and we'll get a single company name here so let's check this one and let's save this so the project execution is over let's quickly move to the project folder to see see the extracted data inside the file so let me open this invoice data file so this is the file and since we have provided the sheet name as the name of the data table that's why we got the data in the data table name that is simple fields and simple Fields formatted we will go with the simple Fields file sheet so this is the invoice number which got extracted this is the due date this is the total amount and this is the company name which got extracted so we got the correct extraction results from the invoices in this file so in this way we can use the combination of machine learning model of AI Fabric and the document understanding package from the UI path to extract data from invoices of different formats this is very helpful since we can extract data from invoices of different form Ms with very less effort and that's all for this session guys I hope you enjoyed this video and if you did give it a like and share with your friends hit the Bell button to get the updates on the latest videos if you have not yet subscribed to our Channel please subscribe to our Channel as well and I'll see you soon in the next one bye-bye