The tests that should be done are about the building, the recognition and the forecast processes.

1)Building: First donwnload Wayeb from here https://github.com/ElAlev/Wayeb. Assuming $WAYEB_HOME is the root directory of Wayeb:

$ cd $WAYEB_HOME

Then build a fat jar:

$ sbt assembly
If it prints a success message, it passes the test.

2) Recognition: In $WAYEB_HOME/data/demo/data.csv you may find a very simple dataset, consisting of 100 events. The event type is either A, B or C. In $WAYEB_HOME/patterns/demo/a_seq_b_or_c.sre you may find a simple complex event definition for the above dataset. It detects an event of type A followed by another event of type B or C. If we want to run this pattern over the stream, we must first compile this pattern into an automaton (make sure you have created a results folder under $WAYEB_HOME):

$ java -jar cef/target/scala-2.12/wayeb-0.2.0-SNAPSHOT.jar compile --patterns:patterns/demo/a_seq_b_or_c.sre --outputFsm results/a_seq_b_or_c.fsm

Now, results/a_seq_b_or_c.fsm is the produced serialized finite state machine. Note that we also provided as input a declarations.sre file. This file simply lets the engine know that the three predicates IsEventTypePredicate(A), IsEventTypePredicate(B) and IsEventTypePredicate(C) are mutually exclusive (i.e., an event can have only one type). This helps the compiler create a more compact automaton. We can use this FSM to perform event recognition on this simple dataset:

$ java -jar cef/target/scala-2.12/wayeb-0.2.0-SNAPSHOT.jar recognition --fsm:results/a_seq_b_or_c.fsm --stream:data/demo/data.csv --statsFile:results/recstats

If it prints information about the throughput and the number of matches, it recognizes the pattern in the stream and it passes the test. 

3) Forecasting: For forecasting, we first need to use a training dataset in order to learn a probabilistic model for the FSM. For this simple guide, we will use $WAYEB_HOME/data/demo/data.csv both as a training and as a test dataset, solely for convenience. Normally, you should use different datasets.

We first run maximum likelihood estimation:

$ java -jar cef/target/scala-2.12/wayeb-0.2.0-SNAPSHOT.jar mle --fsm:results/a_seq_b_or_c.fsm --stream:data/demo/data.csv --outputMc:results/a_seq_b_or_c.mc

The file results/a_seq_b_or_c.mc is the serialized Markov model. The final step is to use the FSM and the Markov model to perform forecasting:

$ java -jar cef/target/scala-2.12/wayeb-0.2.0-SNAPSHOT.jar forecasting --modelType:fmm --fsm:results/a_seq_b_or_c.fsm --mc:results/a_seq_b_or_c.mc --stream:data/demo/data.csv --statsFile:results/forestats --threshold:0.5 --maxSpread:10 --horizon:20 --spreadMethod:classify-nextk

The last command should return some classification statistics like precision, f1 and accuracy.

1) Get familiar with HELIOS instructions
2) Download sample codes and relevant libraries
3) Build the sample APP into a APK (you may modify the sample if you wish)
4) Install the APK to the Android phone (in the absence of a phone, you can use an emulator)
5) Verify with the Android phone that the app launches OK.
6) Send a unique message (such as a random number) to "BugChat" channel in the APP and do the following steps:
  - In the app, open the options menu (the three dots in the right-hand side after title)
  - Tap “Discover others”
  - Search for your nick name and WRITE DOWN the first 6 characters of your ID (it is in the format “nickname @ID”)
7) Fill in the survey and let us know
  - what was the message you sent, with rough time/date information, and
  - the first 6 characters of your ID

Install doc2json

doc2json is located at https://gitlab.ow2.org/decoder/doc_to_asfm

Linux installation

The application requires zlib https://zlib.net/ to retrieve the content of the documents and Clang Tools https://clang.llvm.org/docs/ClangTools.html to convert the user's parsing algorithm into an interruptible reading algorithm.

To install these libraries, you can type the following commands:

> sudo apt-get install zlib1g-dev clang libclang-dev > apt list --installed "libclang*-dev"

If the clang version is less than clang-10 (for instance clang-6), the next cmake build process may fail and you need to update to clang-10 with the following commands

> sudo apt-get install clang-10 libclang-10-dev > sudo apt-get purge --auto-remove libclang-common-6.0-dev > sudo ln -s /usr/bin/clang-10 /usr/bin/clang

You can also check that llvm provides its own header files to the Clang Tools. 

> llvm-config --includedir

should answer a path that contains llvm-xx. Fedora, for instance, returns /usr/include, which prevents the Clang Tools to find some headers like <stddef.h> that are required for string manipulation during the source to source transformation. In such a case, you can try the docker installation that is more robust.

Please note that if after having tested doc2json, you need to revert to your original clang 6 version, just type

# do no use these commands before having built the project and the algorithms libraries > sudo rm /usr/bin/clang > sudo apt-get install clang

Then you can download and build the project

> git clone git@gitlab.ow2.org:decoder/doc_to_asfm.git doc2json > cd doc2json > mkdir build > cd build > cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$PWD .. > make -j 4 # make test is optional = to check if doc2json works on the internal documentation > make test > make install

Docker installation

To download and build it (Docker users) :

> git clone git@gitlab.ow2.org:decoder/doc_to_asfm.git doc2json > cd doc2json > sudo docker build -t doc2json_img . > docker run --name doc2json -it doc2json_img bash # in the docker container > cd /doc2json

Minimal test (to check doc2json works)

Under a Linux system, you can then run some test examples in the build directory, which is also the installation directory, you can type

> ./bin/doc2json ./share/doc2json/libInvoiceReader.so -c ./share/doc2json/config-invoice.xml ../test/invoice.ods -o test/invoice.json > ./bin/doc2json ./share/doc2json/libTextWordReader.so -c ./share/doc2json/config-text-word.xml ../test/StructuredDocument_french.docx -o test/StructuredDocument_french_word.json > ./bin/doc2json ./share/doc2json/libTextOpenofficeReader.so -c ./share/doc2json/config-text-openoffice.xml ../test/StructuredDocument_french.odt -o test/StructuredDocument_french_openoffice.json

Under the Docker container, you can run the same test examples from the /doc2json directory

> ./bin/doc2json ./share/doc2json/libInvoiceReader.so -c ./share/doc2json/config-invoice.xml src/test/invoice.ods -o src/test/invoice.json > ./bin/doc2json ./share/doc2json/libTextWordReader.so -c ./share/doc2json/config-text-word.xml src/test/StructuredDocument_french.docx -o src/test/StructuredDocument_french_word.json > ./bin/doc2json ./share/doc2json/libTextOpenofficeReader.so -c ./share/doc2json/config-text-openoffice.xml src/test/StructuredDocument_french.odt -o src/test/StructuredDocument_french_openoffice.json

to extract the content of the documents in the json format. You can open the document ../test/StructuredDocument_french.docx and compare its content with the result of the extraction, that is the file test/StructuredDocument_french_word.json.

Then, in the build directory for Linux users and in the /doc2json/src directory for Docker users

diff test/StructuredDocument_french_openoffice.json test/StructuredDocument_french_word.json

should show no differences between the two json extracted files even if the origin format (word versus openoffice/opendocument) is very different.

A utility create-reader.sh is provided to generate a parsing library from a custom user's parsing algorithm. Hence the command (in the build directory for Linux users and in the /doc2json directory for docker users)

./bin/create-reader.sh -I . share/doc2json/InvoiceReader.cpp

generates again the parsing library share/doc2json/libInvoiceReader.so from the parsing algorithm share/doc2json/InvoiceReader.cpp. The transformation into a state machine model is generated in the file share/doc2json/InvoiceReaderInstr.cpp.

Apply doc2json to your documents (a text with a title and sections, subsections, subsubsections, ...)

We now suppose that your document is named file.docx or file.odt. It should have a title identified by a specific style.

You may run the test in /tmp with an environment variable DOC2JSON_INSTALL_DIR that refers to the installation directory of doc2json. This is the build directory for Linux users. You can type

export DOC2JSON_INSTALL_DIR=$PWD

in the build directory before trying the test. This environment variable is automatically set for docker users. But you need copy your document into the /tmp directory of the docker container with the command

docker cp .../file.docx doc2json:/tmp

You need to look at the name of the styles the headings of your document.
By default, for a french word document, these styles are Titre, Titre1, Titre2, Titre3, ... For an english word document, they may be Title, Heading1, Heading2, Heading3, ... You need to provide these styles to doc2json by modifying the configuration file config-text-word.xml and by replacing the french Titre by the styles that appear in the style ribbon of Word.

cp $DOC2JSON_INSTALL_DIR/share/doc2json/config-text-word.xml /tmp # edit config-text-word.xml and replace Titre1 by Heading1 or by the styles of the different sections of your document

If the title is not recognized, doc2json will answer "bad generation of the output format!" with a corrupted output json document. This should be modified for a future version.

For the opendocument format, you need to find the style ribbon by clicking on the parameter menu (top right) of libreoffice. Then styles like "Heading 1" should be replaced by "Heading_20_1" in config-text-openoffice.xml - spaces are replaced by "_20_". Sometimes libreoffice renames these styles internally. For instance, it may rename the "Title" style with "P1" or with "P2". The parsing algorithm is not smart enough to recognize this renaming - it will be in a future version. So if the extraction fails, you can unzip your file.odt, and then edit the file content.xml to look for the text of your "Title" and see what is the associated style. Do not spend too much time on that point if it fails.

Then the adequate command

> cd /tmp > $DOC2JSON_INSTALL_DIR/bin/doc2json $DOC2JSON_INSTALL_DIR/share/doc2json/libTextWordReader.so -c /tmp/config-text-word.xml file.docx -o file.json > $DOC2JSON_INSTALL_DIR/bin/doc2json $DOC2JSON_INSTALL_DIR/share/doc2json/libTextOpenofficeReader.so -c /tmp/config-text-openoffice.xml file.odt -o file.json

should extract in file.json the sections and the text of your document.

Feedback questionnaire

When you are done with the testing, please fill in the feedback questionnaire.
Please note that filling in the questionnaire will be your ticket for incentives.

Campaign Mailing List

Please provide your e-mail address below and in the feedback questionnaire, in order to enter the ReachOut incentives programme and to join the mailing list for this campaign, in order to interact with the Campaign Manager. Find out more about Reachout informed consent.

Please read carefully these instructions before completing the Questionnaires.

To connect to the SMOOTH platform and perform the test, please use this link.

The beta version of the software can be accessed at https://www.gotriple.eu  

Instructions:

Test 1: Goal to find which authors published the most on a topic

1. Go to the GoTriple Beta testing platform via the above web address
2. Enter a search term of your choice 
3. Browse the results of the search
4. Select the 'Visual' View of results 
5. Explore the visual view elements
6. Refine the results to show papers from just one of the disciplines provided 
7. Clear the refinement to show results from all disciplines again
8. Find which authors published the most on this topic
9. Click on an author name to view other publications from this author.

Test 2: Goal to produce a Knowledge Map and examine it

1. Make a new search on 'Society + Covid'
2. Refine the results to show only papers published in 2020 
3. Clear the 2020 selection
4. Find book chapters published on this topic (same search Society + Covid)
5. Clear the book chapter selection to return to the overall search list
6. Create a Knowledge Map for this search (be patient it takes a bit of time!) 
7. Examine the knowledge map and see the grouped publications 
8. Return to Home Page

Test 3 Goal: Examine Disciplines and produce a Streamgraph

1. Examine the Disciplines Tab - try clicking on any that are of interest to you
2. View a list of publications from a discipline 
3. Use the filter to refine the results shown
4. Return to the Home page
5. Make a new search on the term 'Co-design'
6. View the Streamgraph for this search   
7. Examine the results of the Streamgraph
8. Visit the GOTRIPLE tab to view project information 

The Toolbox is accessible without the need to log in to the system, but the options are limited to pure search. You can see that without registering the options in the menu are very few. To perform this campaign, we would like all involved users to first sign in into the DataBench Toolbox to get a user profile that you will use throughout the campaign:

- Go to https://databench.ijs.si/ and click on “Sign up” option located at the top of the page on the right side.

- Fill in the form to generate your new user by providing a username and password of your choice, your organization, email, and your user type (at least Technical for this exercise).

Once you have created your user, please sign in with it to the Toolbox. You will be directed to the Toolbox main page again, where you could see that you have more options available. 

Besides the options available through the menu, the main page provides:
A) a carrousel with links,
B) User journeys for users of different profiles: Technical, Business and Benchmarks providers,
C) Videos aimed at these 3 types of users explaining briefly the main functionalities offered for each of them,
D) Shortcuts to some of the of the functionalities, such as FAQ, access to the benchmarks or knowledge catalogues, the DataBench Observatory, etc. 

A) Get information about DataBench pipelines and blueprints

This campaign aims at providing you the means to search and browse existing data pipelines and the explanations on how to map your own architecture to efforts such as the BDV Reference model, the DataBench Framework and the mappings with existing initiatives. 

We encourage you to first go to the Technical user journey  accessible from the front-page of the Toolbox, read it and follow the links given to you to get acquainted with the entries related to blueprints and pipelines. In the “Advanced” user journey you will find the following:

- Link to the DataBench Framework and it relation to the BDV Reference Model, where you can find an introduction to the different elements that composes the DataBench approach towards technical benchmarking.

- Link to the DataBench Generic Pipeline , where an explanation of the 4 main steps in data pipelines are explained. These 4 steps are the basic building blocks for the mappings to other blueprints and existing initiatives.

- User Journey - Generic Big Data Analytics Blueprint : This is the main piece of information that you need to understand what we mean by mapping an existing architecture to our pipelines and blueprints. You will find links to the generic pipeline figure.

- Practical example of creating a blueprint and derived cost-effectiveness analysis: Targeting the Telecommunications Industry .

- Ways to report your suggestions for new blueprints, by using the Suggest blueprint/pipeline option  under the Knowledge Nuggets menu 

Below is a summary of the minimal set of actions we encourage you to do:

1. Go to the User journeys area of the main page and click on “Technical”.
   
   2. Go to the link to the User Journey: Generic Big Data Analytics Blueprint  at the bottom of the “Advanced” area of the page. 

3. Read and understand the different elements of the pipeline (the 4 steps) and the elements of the generic blueprint as described in the previous link.

4. Check examples of already existing blueprints. In order to do that use the search box located at the top right corner and type “blueprint”. Browse through the blueprints. 

B) Desk analysis

Once you are familiar with the DataBench Toolbox and the main concepts related to the blueprints, you need to do some homework. You should try to map your own architecture to the DataBench pipeline and the generic blueprint. We suggest the following steps:

- Prepare a figure with the architecture you have in mind in your project/organization. 

- Create links to the 4 steps of the data pipeline and generate a new figure showing the mapping.

- Create links to the Generic Big Data Analytics Blueprint  figure and generate a new figure showing the mappings. In order to do so you might use the generic pipeline figure and particularize to your components as it was done in the example provided for the Telecommunications Industry . 

C) Upload your blueprint to the Toolbox

- Upload your files as pdf or images by using the Form of suggestion of blueprints   available from the Knowledge Nuggets menu. Try to include a description with a few words about the sector of application of your blueprint, main technical decisions or anything you might find interesting to share. 

- The DataBench project will revise the blueprints and publish them into the platform acknowledging your authorship. 

Congratulations! You have completed the assignment of this campaign! Go now to fill in the feedback questionnaire. Please note that filling in the questionnaire will be your ticket for incentives.

The Toolbox is accessible without the need to log in to the system, but the options are limited to pure search. You can see that without registering the options in the menu are very few. 

Initial steps to log in as a Toolbox user

To perform this campaign, we would like all involved users to first sign in into the DataBench Toolbox and create  a user profile that you will use throughout the campaign:

- Go to http://databench.ijs.si/ and click on “Sign up” option located at the top of the page on the right side.
- Fill in the form to generate your new user by providing an username and password of your choice, your organization, email, and your user type (Technical and/or Business, depending on your preferences and skills).

Once you have created your user, please sign in with it into the Toolbox. You will be directed to the Toolbox main page again, where you could check that you have more options available. 

Besides the options available through the menu, the main page provides:
A) a carrousel with links,
B) User journeys for users of different profiles: Technical, Business and Benchmarks providers,
C) Videos aimed at these 3 types of users explaining briefly the main functionalities offered for each of them,
D) Shortcuts to some of the functionalities, such as FAQ, access to the benchmarks or knowledge catalogues, the DataBench Observatory, etc. 

A) For Technical Users

This campaign aims at using the user journeys as starting point to help you navigating the tool. We encourage you to click on the Technical user journey, read it and follow the provided links to get acquainted with the tool and what you can do with it. Get used to the main two catalogues: the benchmarks catalogue (tools for big data and AI benchmarking), and the knowledge nuggets catalogue (providing information about technical and business aspects related to benchmarking and big data technologies). Learn about existing big data architectural blueprints and browse some of them. 

Additionally, if you already have a goal in your mind (i.e. finding a benchmark for testing a specific ML model, or compare the characteristics of different NoSQL databases), we encourage you to try to find the appropriate benchmark and report your conclusions later in the questionnaire. 

Below is a summary of the minimal set of actions we encourage you to do:

1. Go to the User journeys area of the main page and click on “Technical”. 

2. Read the content of this page, divided into advice for “Beginners” (first-time users) and “Advanced” (providing extra recommendations of what to do next). Focus first on the “Beginners” area and click on the different links to browse to the different options to get used to the tool. We recommend you to come back to the User journey page  until you click on all the available options for beginners, but feel free to stray and use the navigation and links from other pages to get used to the tool. After you finish clicking on all the options for beginners, you should have seen the benchmarks and knowledge nuggets catalogues, used some of the search functionalities and browsed some of the existing architectural blueprints. You are now ready to go further!

3. Focus now on the “Advanced” area of the User journey page . 

- Here you will find ways to suggest new content via web forms (i.e. new benchmarks you might know that are missing in the catalogue, a version of a big data blueprint you are dealing with in a project, or a new knowledge nugget based on your experience). We are not expecting you to fill-in these forms at this stage, but just acknowledge their potential value (and feel free to contribute any time).  

- You will find also links to specific more advanced user journeys or practical examples at the end of the advanced user journeys. Click on the ones that take your attention and start navigating via the links offered by them. From this moment we expect that you know the main options of the Toolbox and how to navigate and browse through it. You should have noted by now that both benchmarks and knowledge nuggets are annotated or categorized with clickable tags, which makes navigation through related items possible.  

4. Get used to the search functionalities. The Toolbox offers 4 types of search:
- Search text box located at the top right corner of the pages. This is a full text search. You can enter any text and the results matching that text from both the benchmark and knowledge nuggets catalogues will appear.

- Search by “BDV Reference Model”  option from the menu allows you to have a look at the model created by the BDV PPP community (check the BDV SRIA  for more details). The model is represented graphically and is clickable. If you click in any of the vertical or horizontal layers of the model you will be directed to the benchmarks and/or knowledge annotated in the Toolbox to these layers. Browse through this search.

- Search by “Guided benchmark search” . In simple terms this is a search by the tags used to annotate benchmarks and knowledge nuggets. These tags range from technical to business aspects. You can click on the categories of tags to find related information. Browse to some of the options of this search.  

- Finally, the “Search by Blueprint/Pipeline”  option allows a search that presents graphically a generic architectural blueprint developed in DataBench with the most common elements of a big data architecture. The blueprint is aligned with 4 steps of a DataBench Generic data pipeline  (data acquisition, preparation, analysis and visualization/interaction). The graphic is clickable both at the level of the four steps of the pipeline and in some of the detailed elements of the blueprint. Click on the parts of the diagram you are interested to find a list of existing benchmarks and nuggets related to it. Browse some of them. There are nuggets that show a summary of existing big data tools for each of the elements of the pipeline. See if you find it easy to browse through the results .

Congratulations! You have completed the assignment of this campaign! Go now to fill in the feedback questionnaire. 

NOTE – Some of the available benchmarks can be deployed and run in your premises. Those are listed first in the Benchmark catalogue and when you click on them you will find the configuration file at the bottom of their description. If you want to run any of them, you should have dedicated infrastructure to do so. We are not expecting you to do so in this exercise.

B) For Business users

As for technical users, this campaign aims at using the user journeys as starting point to help you navigating the tool. We encourage you to click on the Business user journey, read it and follow the links given to you to get acquainted with the tool and what you can do with it. Get used to the main two catalogues: the benchmarks catalogue (tools for big data and AI benchmarking), but mainly to the knowledge nuggets catalogue (providing information about technical and business aspects related to benchmarking and big data technologies). Learn about existing big data architectural blueprints and browse to some of them, as they apply to different industries and might be of interest for business purposes.

Additionally, if you already have a goal in your mind (i.e. finding most widely used business KPIs in a specific sector), we encourage you to try to find the appropriate information in the knowledge nugget catalogue and report your conclusions later in the questionnaire. 

Below there is a summary of the minimal set of actions we encourage you to do:

1. Go to the User journeys area of the main page and click on “Business”. 

2. Read the content of this page, divided into advice for “Beginners” (first-time users) and “Advanced” (providing extra recommendations for what to do next). Focus first on the “Beginners” area and click on the different links to browse to the different options to get used to the tool. We recommend you to come back to this User journey page  until you click on all the available options for beginners, but feel free to stray and use the navigation and links from other pages to get used to the tool. After finishing clicking on all the options for beginners, you should have seen the benchmarks and knowledge nuggets catalogues, used some of the search functionalities and browsed some of the existing architectural blueprints. You are now ready to go further!

3. Focus now on the “Advanced” area of the User journey page .
- You will find links to different elements, such as nuggets related to business KPIs, by industry, etc. Browse through them and follow the links.

- You will find ways to suggest new content via web forms (i.e. a new knowledge nugget based on your experience). We are not expecting you to fill-in these forms at this stage, but just acknowledge their potential value (but feel free to contribute any time).  

- You will find also links to specific more advanced user journeys or practical examples at the end of the advanced user journeys. Click on the ones that take your attention and start navigating via the links offered by them. From this moment we expect that you know the main options of the Toolbox and how to navigate and browse through it. You should have noted by now that both benchmarks and knowledge nuggets are annotated or categorized with clickable tags, which makes navigation through related items possible.  

5. Get used to the search functionalities. The Toolbox offers 4 types of search:
- Search text box located at the top right corner of the pages. This is a full text search. You can enter any text and the results, matching that text from both the benchmark and knowledge nuggets catalogues, will appear.

- Search by “BDV Reference Model”  option from the menu allows you to have a look at the model created by the BDV PPP community (check the BDV SRIA  for more details). The model is represented graphically and is clickable. If you click on any of the vertical or horizontal layers of the model you will be directed to the benchmarks and/or knowledge annotated in the Toolbox to these layers. Browse through this search.

- Search by “Guided benchmark search” . In simple terms this is a search by the tags used to annotate benchmarks and knowledge nuggets. These tags range from technical to business aspects. You can click on the categories of tags to find related information. Browse to some of the options of this search.  

Finally, the “Search by Blueprint/Pipeline”  option allows a search that presents graphically a generic architectural blueprint developed in DataBench with the most common elements of a big data architecture. The blueprint is aligned with 4 steps of a DataBench Generic data pipeline  (data acquisition, preparation, analysis and visualization/interaction). The graphic is clickable both at the level of the four steps of the pipeline and in some of the detailed elements of the blueprint. Click on the parts of the diagram you are interested to find a list of existing benchmarks and nuggets related to it. Browse some of them. There are nuggets that show a summary of existing big data tools for each of the elements of the pipeline. See if you find them it easy to browse through the results .
6. This part of the test is not guided, as we expect you to navigate through the options you have seen previously. Once you know how to navigate, try to find information of interest for your industry or area of interest:
• Try to find information about the most widely used KPIs or interesting use cases.
• Try to find information about architectural blueprints for your inspiration.  

Congratulations! You have completed the assignment of this campaign! Go now to fill in the feedback questionnaire.