diff --git a/user_guide/01-B-detailed-installation.Rmd b/user_guide/01-B-detailed-installation.Rmd index 3251157..aac9292 100644 --- a/user_guide/01-B-detailed-installation.Rmd +++ b/user_guide/01-B-detailed-installation.Rmd @@ -3,7 +3,8 @@ Selecting an Installation Strategy ---------------------------------- -When you install VisionEval, you need to choose whether do do an [End-User Version](#enduser-installation) or to set up a full [Development Environment](#development-environment)? +When you install VisionEval, you need to choose whether to perform an [End-User Installation](#enduser-installation) +or to set up a full [Development Environment](#development-environment)? Here are some considerations: @@ -13,9 +14,10 @@ Here are some considerations: - You can install both and go back and forth if you like (especially if you create a [Production Environment](#production-environment)). Setting up a [Production Environment](#production-environment) will let you keep all your models in their own folder -separate from where the VisionEval code resides. Use a Production Environment to avoid having to move your -models when you upgrade to a new VisionEval version. A Productrion Environment also makes it easier to keep your models -backed up without having to copy gigabytes of VisionEval code, or to create a Github repository just for your model. +separate from where the VisionEval code resides. Use a Production Environment to avoid having to move your models when +you upgrade to a new version of VisionEval or R. A Production Environment also makes it easier to keep your models +backed up (without having to copy gigabytes of VisionEval code), and straightfoward to create a Github repository just +for your model. ### Recommended installation workflow: @@ -25,13 +27,14 @@ backed up without having to copy gigabytes of VisionEval code, or to create a Gi - Start by connecting the Production Environment to the End User Installation 3. Add a [Development Installation](#development-environment) to build VisionEval on your own - You can just do "Development Light" ('`src`' installation within your End User installation) if you are re-creating certain packages + ([Instructions Here](#ve-buildprocess)) - Or you can do a Full Development Version (Github) - The Full Development Version lets you stay current with the most recent VisionEval development. - You'll need to build the Full Development Version locally before you can run your model. - 4. Re-connect your Production Environment to the Development Installation when you're ready + - Re-connect your Production Environment to the Development Installation when you're ready 5. When you get a new version of VisionEval, update the links in your Production environment - - `VE_RUNTIME` in your VisionEval Home `.Renviron` (i.e. where VisionEval code is) - - `VE_HOME` in your VisionEval Runtime `.Renviron` (i.e. where your VisionEval models are) + - `VE_RUNTIME` to `.Renviron` in your VisionEval Home (i.e. where VisionEval code is) + - `VE_HOME` to `.Renviron` in your VisionEval Runtime (i.e. where your VisionEval models are) Full instructions are found below. @@ -43,13 +46,14 @@ Full instructions are found below. - Pick a supported R version - If you already have R installed, make sure VisionEval is available for your R version. - You can have multiple R versions installed on your computer. Just make sure you use one that is consistent with the VisionEval code you have installed. - - Choose R 4.3.2 if you haven't already chosen for other reasons - - You can download a specific supported version of R (as of January, 2024) using the following links: + - Choose the latest suppored R (R 4.3.2 as of January 2024) if you haven't already chosen for other reasons + - You can download a specific supported version of R (as of January 2024) using the following links: - R 4.1.3: [https://cran.r-project.org/bin/windows/base/old/4.1.3](https://cran.r-project.org/bin/windows/base/old/4.1.3) - R 4.2.3: [https://cran.r-project.org/bin/windows/base/old/4.2.3](https://cran.r-project.org/bin/windows/base/old/4.2.3) - R 4.3.2: [https://cran.r-project.org/bin/windows/base/old/4.3.2](https://cran.r-project.org/bin/windows/base/old/4.3.2) -1. Install RStudio. You don't need RStudio to run VisionEval, but many people find it helpful. If you computer is locked down, you won't be able to install it as it requires administrative permissions. +1. Install RStudio. + - You don't need RStudio to run VisionEval, but many people find it helpful. If your computer is locked down, you won't be able to install it as it requires administrative permissions. - Get the latest version here: [https://posit.co/download/rstudio-desktop/](https://posit.co/download/rstudio-desktop/) - RStudio is now (November 2022) being maintained and distributed by a new company called "Posit" (the same key people are involved) - Start RStudio and check that it is using your intended version of R. The startup message in the console window will show the R version, or you can display the built-in variable `R.version` and look at the `version.string` @@ -90,27 +94,27 @@ Full instructions are found below. ------------------------------------------------------------------------ Though it is possible to build your own local VisionEval model in the "models" folder of your VisionEval installation, -we don't recommend that since you will need to move the model every time you install a new version of VisionEval. Also, -having your models in their own location makes it much easier to back up the models or to put them under version -control without having to bring along the entire VisionEval code base as well. +we don't recommend that since you will need to move the model every time you install a new version of VisionEval. +Having your models in their own location makes it much easier to back up the models or to put them under version +control without having to bring along the entire VisionEval code base. -Instead, you can set up a "Production Environment" and store your model in a location distinct from where the VisionEval -code is located. That way you can upgrade (or downgrade) VisionEval without having to move your model, or you can shift -from using the end-user installation (downloadable .zip installer) to a development installation (cloning the Github -repository). +To keep the code and models separate, you can set up a "Production Environment" and store your model in a location +distinct from where the VisionEval code is located. That way you can upgrade (or downgrade) VisionEval without having to +move your model, or you can shift from using the end-user installation (downloadable .zip installer) to a development +installation (cloning the Github repository). By default, VisionEval looks for models in the "models" folder of the VisionEval runtime, which itself defaults to the folder into which you extracted the VisionEval installer .zip file. -- In production, that's a problem because you will need to move the model setups every time you change +- In production, the default behavior is a problem because you will need to move the model setups every time you change the version of R or VisionEval -- You can save headaches by setting up a separate permanent folder for your models, which we'll refer to as `VE_RUNTIME` +- You can prevent headaches by setting up a separate permanent folder for your models, which we'll refer to as `VE_RUNTIME` The VisionEval startup code knows about two locations referred to in environment variables (see below if you don't know -what those are): +what environment variables are): -- `VE_HOME`, which is where the VisionEval application code is located -- `VE_RUNTIME`, which is where the "models" folder is located +- `VE_HOME` is where the VisionEval application code is located +- `VE_RUNTIME` is where the "models" folder is located These can be entirely different folders on different drives. Out of the box, `VE_HOME` and VE_RUNTIME are both the same, and they point to the folder where you unzipped VisionEval. @@ -119,14 +123,14 @@ By setting `VE_HOME` and `VE_RUNTIME` yourself, you can start visioneval from th simply adjusting `VE_HOME` will allow you to run the model with a different version of R or VisionEval. You can set VisionEval up to start either from within `VE_HOME` or from within `VE_RUNTIME. Separating these two -location is what we refer to here as creating a "**Production Environment**". +locations is what we call "creating a **Production Environment**". We suggest that you set up your system so you can start VisionEval from `VE_RUNTIME` (the place you store your models), since it is then easier to update your R or VisionEval version. Here is how to set that up: - The first thing to do is decide where you want your VisionEval runtime (`VE_RUNTIME`) - Create that folder (for example `%USERPROFILE%/VE_Models`) -- Go to your VisionEval installation (`VE_HOME` where you installed the VisionEval code) +- Go to your VisionEval installation (`VE_HOME` which is the place where you installed the VisionEval code) - Open the `.Renviron` file with your favorite text editor (or Notepad, which is nobody's favorite, but which will do the job) - If the `.Renviron` file does not exist, please start your VisionEval installation once using VisionEval.Rproj or launch.bat and @@ -139,15 +143,13 @@ since it is then easier to update your R or VisionEval version. Here is how to s ``` Now when you start VisionEval by clicking `VisionEval.Rproj` (or `launch.bat`), you'll see that it is running in the -folder you assigned to `VE_RUNTIME`. Also notice that an empty "models" directory created when you ran VisionEval for -the first time. That "models" folder is where you will be creating and running your own VisionEval models. +folder you assigned to `VE_RUNTIME`. Also notice that an empty "models" directory was created there when you ran +VisionEval for the first time. The "models" folder is where you will be creating and running your own VisionEval +models. To start VisionEval from within your `VE_RUNTIME` folder, you will need to copy some of the runtime support files from the installation location. -To be able to start VisionEval from within your `VE_RUNTIME`, you need to copy some files from the installation -runtime and adjust a line in the `.Renviron` file. - - Copy these runtime files from your installation directory to your VE_RUNTIME: - `launch.bat` - `.Rprofile` @@ -164,32 +166,38 @@ The `.Renviron` file should already have `VE_HOME` correctly defined, but you ca You can also define `VE_HOME` yourself. It should point to the folder where you unzipped the VisionEval installer, or to the folder into which you cloned the Github repository. -If you later change the R or VisionEval version, you can just copy over the new .Renviron or manually adjust the `VE_HOME` definition with a text editor. +If you later change the R or VisionEval version, you can just copy over the `.Renviron` file from the new installation after +you have opened it once in R (or you can manually edit `.Renviron` to point to the new `VE_HOME` location. To configure your `VE_RUNTIME` to use a development environment (the cloned Github repository) the steps are a little different: - Build visioneval (see the section on setting up a development environment and building VisionEval, below). That step - is necessary to have runnable VisionEval code, and for it to generate a suitable .Renviron file for locating + is necessary to have runnable VisionEval code, and for it to generate a suitable `.Renviron` file for locating `VE_HOME`. -- Locate and copy the runtime files from "MyVisionEval-Github-clone/built/visioneval/4.3.2/runtime" (or whatever R - version you built VisionEval with). If you set the `VE_BUILD` variable (see below for setting up a development - environment), the runtime files will be located in `VE_BUILD/built/VE_BRANCH/VE_R_VERSION/runtime` where: - - VE_BUILD is the directory you set up to receive the built - - VE_BRANCH is the Github branch you are building from. The default for `ve.build` is "visioneval" and all branches will - be built into that same location). - - You can create different builds for the Github branches by running `ve.build(use.git=TRUE)` - - VE_R_VERSION is the version of R with which you built VisionEval. +- Find and copy the runtime files from "MyVisionEval-Github-clone/built/visioneval/4.3.2/runtime" (or whatever R + version you built VisionEval with). These are the files to copy: + - `launch.bat` + - `.Rprofile` + - `.Renviron` + - `VisionEval.Rproj` +- If you set the `VE_BUILD` variable (see below for setting up a development environment), the runtime files will be located + in `VE_BUILD/built/VE_BRANCH/VE_R_VERSION/runtime` where: + - VE_BUILD is the directory you set up to receive the built + - VE_BRANCH is the Github branch you are building from. The default for `ve.build` is "visioneval" and all branches will + be built into that same location. + - You can create different builds for the Github branches by running `ve.build(use.git=TRUE)` + - VE_R_VERSION is the version of R with which you built VisionEval. - Do NOT delete the original runtime folder from the `built` directory (the startup script requires it to be present, even if it is not going to be used). - Finally, find the generated `.Renviron` file in the root directory of your development environment and copy that to `VE_RUNTIME`. - `VE_HOME` in `.Renviron` should be the location of the Git root directory (NOT the built runtime) - It will still work if you point `VE_HOME` at the built runtime (where the runtime files are located), you just won't have access to the developmment environment tools. - - The Github root also contains a `launch.bat` and a `VisionEval.Rproj`. These identical to the ones it the runtime - and you can copy them to your `VE_RUNTIME` - - To start VisionEval from your development environment (Github clone), you will need to add (or edit) `VE_RUNTIME` - in the `.Renviron` file located in the development environment root directory. + - The Github root also contains a `launch.bat` and a `VisionEval-dev.Rproj`. You can use those to start VisionEval. + - To start VisionEval from your development environment (Github clone), you will need to add (or edit) the `VE_RUNTIME` + line in the `.Renviron` file located in the development environment root directory (set it to the full path + of your runtime folder). Setting up a Development Environment ------------------------------------------------------------------------ @@ -197,7 +205,9 @@ different: A development environment will allow you to build and rebuild VisionEval from the Github. Building VisionEval is easy; it requires a little more software setup and it's just a bit time-consuming the first time. Here are the basic steps: - - You need one additional software installation (RTools43), in addition to R and (optionally) RStudio + - You will need a copy of git to clone the development source code, or you can download a snapshot of the code as a + zip file from Github + - You need one additional software installation (RTools43), in addition to R and (optionally) RStudio to run VisionEval - Then you just start R from the root of the Github and run the `ve.build()` function After that you can make any change you want to VisionEval and rebuild those changes just by running `ve.build()` again. @@ -210,6 +220,9 @@ To get the development code, you clone the Github repository.You can find Vision development version, which is probably what you want to use if you're doing any development rather than just installing. +If you don't want to work with git, but would still like to build from code, you can do that. Visit the [Github](https://github.com/VisionEval/VisionEval-Dev), +nd hit the green Code button, select "Local", then choose "Download ZIP". + If you plan to contribute code back to the VisionEval project, you should work with `Visioneval-dev`. If you're just making local changes or rebuilding certain packages for your own use, you can use `VisionEval` @@ -221,7 +234,7 @@ can use the [Github desktop client](https://desktop.github.com). Either of those To clone the repository, you will want to execute the following instruction in your Git client of choice. Please be aware that you do *NOT* need to "fork" the repository in order to make a clone and build VisionEval locally. See the -discussion of forking below for more information. +discussion of forking below for more information. Most people will NOT want to "fork". ``` cd /folder/for/Git-Repos @@ -230,7 +243,7 @@ git clone https://github.com/visioneval/VisionEval-dev.git VisionEval-dev The rest of this section on the development environment will presume your VisionEval git clone is in a folder called "VisionEval-dev". If you're attaching your [Production Environment](#production-environment) to your new VisionEval -clone, use the Github root folder (e.g. `/folder/for/Get-Repos/VisionEval-dev`) as the value for the `VE_HOME` +clone, use the Github root folder (e.g. `/folder/for/Git-Repos/VisionEval-dev`) as the value for the `VE_HOME` environment variable. Or you can just copy the `.Renviron` file that is created when you first run `ve.build()`, which will set VE_HOME to the correct value. @@ -274,11 +287,11 @@ When you run `ve.build()` it creates two new non-versioned folders in "VisionEva - In the existing build directory, a build configuration file is created called "ve-output-visioneval-4.x.x.make" where the "4.x.x" is whatever R version you were building for - Two new folders are created at the top of the repository: - "dev" which contains build log files, build status, and a library of R packages that are used during building but are not included in the end-user installation - - "built" which contains the elements of the build VisionEval + - "built" which contains the built compoenents - The directory structure under "built" looks like this: - built - - visioneval (or really, VE_BRANCH - see below to set up building different git branches separately) - - 4.x.x (where x.x is whatever sub-version of R you're using; if you build a different R version from the same repo you'll get additional folders) + - visioneval (or really, VE_BRANCH - see below to build different git branches separately) + - 4.x.x (where x.x is whatever sub-version of R you're using; if you build a different R version from the same repository branch you'll get additional folders) - docs (internal documentation, which will be empty unless you do `ve.build("docs")` - pkg-dependencies (a local R repository with all the dependency packages needed by VisionEval; these are also installed into ve-lib below) - pkg-ve-repo (a local R repository with the built-but-uninstalled VisionEval packages) @@ -289,7 +302,7 @@ When you run `ve.build()` it creates two new non-versioned folders in "VisionEva - ve-pkg (installable tar.gz package files that are only built for Mac or Linux for a source installation on those platforms; usually this folder will be empty on Windows though there are magic internal flags that can force ve-pkg to be populated) You should NOT put real work or real models into the "runtime" or "runtime.test" folders. Those will get blown away when -you rebuild VisionEval. Set up a [Production Environment](#production-environment) instead +you rebuild VisionEval. Set up a [Production Environment](#production-environment) instead to define `VE_RUNTIME`. If you have a git feature branch for VisionEval-dev, you can build that branch separately from the default git branch (and if that makes no sense to you, you don't need to know...). @@ -310,8 +323,8 @@ here), it can help to put all your VisionEval builds in the same separate folder environment" similar to your "production environment". To put your builds somewhere else than in your git repository clone, just define a `VE_BUILD` environment variable in the -.Renviron in the root of your development environment (VisionEval-dev). `VE_BUILD` should name an existing folder -.somewhere outside your VisionEval git directory (e.g. :%USERPROFILE%\VisionEval-build") +`.Renviron` in the root of your development environment (VisionEval-dev). `VE_BUILD` should name an existing folder +somewhere outside your VisionEval git directory (e.g. :%USERPROFILE%\VisionEval-build") Once you have defined VE_BUILD, restart R or RStudio, and then when you run `ve.build()` or the command line equivalent, you'll get this structure of files and folders: @@ -328,6 +341,8 @@ you'll get this structure of files and folders: `ve.build()` will still create ".Renviron" file in the git root and also create the "build/ve-output-visioneval-4.x.x.make" file within the Git repository +If you define `VE_BUILD`, you do NOT need to change `VE_HOME`: it should still point to the root of your VisionEval git repository. + ### Frequently Asked Questions (development) - Do I need to "fork" the VisionEval-dev repository (short answer: "NO") and why would I ever want to? @@ -347,8 +362,8 @@ you'll get this structure of files and folders: - If you are doing "pull requests", they should always be made against the "development" branch of VisionEval-dev, and you should be sure to rebase your changes on the current HEAD of the "development" branch prior to submitting the pull - request (or it is guaranteed to be rejected). (If none of that makes any sense to you, then you shouldn't even be - thinking of forking or making a pull request.) + request (or it is guaranteed to be rejected). If none of that makes any sense to you, then you shouldn't even be + thinking of forking or making a pull request. - What if I'm not using Windows? @@ -359,13 +374,13 @@ you'll get this structure of files and folders: depending on the speed of your machine). - You will need a LOT of RAM to do the build (we recommend 16 Gigabytes) so if you're trying this on a Linux cloud server, you'll want a premium setup configured. Once you've built it, it will run with less RAM (or work - reasonably with a large swap file). Remember that running a large VE model may need as much as 8 or 9 Gigabytes of + reasonably with a large swap file). Remember that running a large VE model may still need as much as 8 or 9 Gigabytes of RAM per concurrent scenario. - You do not need RStudio on Mac or Linux (but you could use it, if your system is [supported by RStudio](https://posit.co/download/rstudio-desktop/) and if the system has a graphical user interface). - - You do NOT need RTools, which is only for Windows. The RTools functions are part of the standard Mac or Linux OS - environments. On Linux, depending on your distribution, you may need to install the system development environment - package (furnishing make, GNU compilers, etc) as well as the R development package (not just base R).. + - You do NOT need RTools, which is only for Windows. The RTools functions are part of the standard Mac or Linux operating system + environments. On Linux, depending on your distribution, you will need to install the system build environment (e.g. the "build-essential" package on Debian + or ubuntu). That will furnish make, GNU compilers, etc. You will also need the full R packages (r-base, and r-base-dev). - Without RStudio, just follow the command line instructions below - *Caveat* No one has recently (1/2024) tried a non-Windows build, so there may still be some weirdnesses (e.g. with file permissions). Contact jeremy.raw at dot.gov if you're having trouble with a non-Windows build. @@ -376,28 +391,34 @@ you'll get this structure of files and folders: Bash command line, we're not going to explain it here. Just use RStudio or launch.bat (for the standard R GUI) and `ve.build()`. - Here's the command line drill: - Open the RTools43 Bash command line (it's an icon within the RTools43 Start Menu) - - Change into the `build` folder of your git repository clone (that is NOT the same as the `built` folder - `build` - contains the code needed to perform the build operations. + - Change into the `build` folder of your git repository clone. That is NOT the same as the `built` folder - `build` + contains the code needed to perform the build operations, and it is part of the versioned Github code. - Type `make` and then go do something else for an hour or so. - When `make` is done, go back to the git root in Windows Explorer and double-click launch.bat (or VisionEval-dev.Rproj) to start the built VisionEval with "ve.run()" - - If you want to run a command line inside RStudio, you can open a Terminal window, change into the "build" directory, - and run "make" from there. That's it! Everything else is just like doing `ve.build()` from within RStudio or the R - GUI, though there are some more options to do partial builds. See `Makefile.md` in the `build` directory to learn - about all the build options. + - If you really love the terminal, you can put the R executable on your PATH and then after `make` finishes, just change to + the root of your git installation. Then enter "R" and you'll get an R command line right there with VisionEval running + within it. But of courses, who would want to do that if they weren't setting up a cluster of machines to do massively + parallel VisionEval scenario runs? + - If you want to run a command line inside RStudio, you can open a Terminal window, change into the "build" directory, + and run "make" from there. That's it! Everything else is just like doing `ve.build()` from within RStudio or the R + GUI, though there are some more options to do partial builds. + See [`Makefile.md`](https://github.com/VisionEval/VisionEval-Dev/blob/development/build/Makefile.md) in the `build` + directory to learn about all the build options. By default, the command line "build" process does not save a log of what it did. (If you run `ve.build()` from within - R, the console window keeps the log). Here's a magic incantation for the RTools43 Bash commend line that will save a - log file (in "build/make.out" - you can use any other file name or location your like): + R, the console window keeps the log). Here's a magic incantation for the RTools43 Bash commend line (or Mac and Linux terminals) + that will save a log file (in "build/make.out" - you can use any other file name or location your like): ``` nohup make >make.out 2>&1 & tail -f make.out ``` - Type `Ctrl-C` to exit the `tail` program (which shows the log in the terminal window as the build runs). + Type `Ctrl-C` to exit the `tail` program (which shows the log in the terminal window as the build runs). Start monitoring + again by re-entering just the `tail -f make.out` part. If the `tail` display doesn't change for a while, stop it with + `Ctrl-C` and see if you get the "done" message. - If you then want to end the `make` process, end the tail command (with `Ctrl-C`) and then enter this commend + If you then want to end the `make` process, end the tail command (with `Ctrl-C`) first then enter this commend (which gets the make process ID and stops the process): ``` diff --git a/user_guide/10-build-process.Rmd b/user_guide/10-build-process.Rmd index fdcdab4..2c18e11 100644 --- a/user_guide/10-build-process.Rmd +++ b/user_guide/10-build-process.Rmd @@ -22,7 +22,7 @@ VisionEval relies on this fundamental package framework for sharing code and dat #### Build vs Runtime Environment -It is crucial for users to understand the distinction between the VisionEval runtime environment and the VisionEval build environment. The VisionEval begins as source code and data in the build environment from which the VisionEval R packages are built, and a runtime environment is created. VisionEval models can only be run from a runtime environment where the built packages are loaded (into an R session) and used by R to perform the model steps. VisionEval runtime environment will be referred to as *VisionEval* from this point on in the document. +It is crucial for users to understand the distinction between the VisionEval runtime environment and the VisionEval build environment. VisionEval begins as source code and data in the build environment from which the VisionEval R packages are built, and a runtime environment is created. VisionEval models can only be run from a runtime environment where the built packages are loaded (into an R session) and used by R to perform the model steps. The VisionEval runtime environment will be referred to as *VisionEval* from this point on in the document. ## Where do we Start the ‘Re-Building’ Process @@ -40,27 +40,19 @@ Users at this point should have a full *VisionEval* installed and running. If no #### Software Requirements -- *R* – R for windows can be downloaded here: *(Tested using 4.1.3). It is suggested to use an R version that aligns with the VE R version. The Cran website includes links to past versions although the most current is the one often highlighted on the first page.* +- *R* – R for windows can be downloaded here: *(The most recent version is usually supported in VisionEval within a few weeks of its release). It is suggested to use an R version that aligns with the VE R version. The Cran website includes links to past versions although the most current is the one often highlighted on the first page.* -- *RStudio –* RStudio for windows can be downloaded here: +- *RStudio –* RStudio for windows can be downloaded here: #### (Re)Building Package -Rebuilding a package in this *VisionEval* setup will require source code for the specific build that was installed (specific R, operating system, etc.) which is made available in a zip file \[here\]. This zip file contains the source code and default data for all the VisionEval packages that can be easily installed in *VisionEval* using R native command “install.packages” or [RStudio build tools](https://support.posit.co/hc/en-us/articles/200486518-Customizing-Package-Build-Options-in-the-RStudio-IDE). The user can download and unzip the zip file into their *VisionEval* directory. After unzipping, the *VisionEval* directory structure should look similar to Figure 1. +Rebuilding a package in this *VisionEval* setup will require source code for the specific build that was installed (specific R, operating system, etc.) which is made available in a zip file that you can get from [the Github release](https://github.com/VisionEval/VisionEval-Dev/releases). Be sure to pick the version corresponding your installed version of R and VisionEval. This zip file contains the source code and default data for all the VisionEval packages that can be easily installed in *VisionEval* using R native command “install.packages” or [RStudio build tools](https://support.posit.co/hc/en-us/articles/200486518-Customizing-Package-Build-Options-in-the-RStudio-IDE). The user can download and unzip the zip file into their *VisionEval* directory. After unzipping a `src` folder will be present alongside `ve-lib`. Each VisionEval package is present in a sub-folder of `src` and you can change files there and rebuild the package as described below. -![](images/10_image1.png) - -Figure 1 VisionEval runtime directory structure with source code unzipped - -The source code of these packages contains code and data upon which the modules are built. This data can then be updated with localized or custom data and the module can be rebuilt into *VisionEval*. The process of rebuilding is explained in more detail in the \[Common Use Cases\] section of this document. +The source code of these packages contains code and data upon which the modules are built. This data can then be updated with localized or custom data and the module can be rebuilt into *VisionEval*. The process of rebuilding is explained in more detail in the section of this document. ### Users who created *VisionEval* through source code from GitHub -For advanced users who have a complete *VisionEval* built using the source code from GitHub, the process for updating and rebuilding individual modules and packages can be done without the need to download additional source code material. A version of the source code, easily installable in R, is created during the “build process”. A default directory structure for this setup is shown in Figure 2. - -![Directory structure for built VisionEval](images/10_image2.png) - -Figure 2 VisionEval build directory structure +For advanced users who have a complete *VisionEval* built using the source code from GitHub, the process for updating and rebuilding individual modules and packages can be done without the need to download additional source code material. A version of the source code, easily installable in R, is created during the “build process”. *Note: if you want to install a complete VisionEval model from the source code please see the section* *VisionEval Full Build Process.* @@ -68,7 +60,7 @@ Figure 2 VisionEval build directory structure - *RTools* – RTools for windows can be downloaded here: . -- *Git* – Git for windows can be downloaded here: . +- *Git* – Git for windows can be downloaded here: . #### (Re)Building Packages Using Package Sources @@ -88,11 +80,7 @@ Advanced users may want to build VisionEval natively from the source code for de - For developers, there is a separate development repository with the latest unreleased source code: - - Users not familiar with [git](https://git-scm.com/downloads) and GitHub can simply download the repository as a zip file (). In either case, the download is very large (over 900MB compressed download and 3.4GB uncompressed on disk). The Zip file download is shown in the figure below. - -   - - - [![](images/image3.png "fig:")](https://github.com/VisionEval/VisionEval/archive/refs/heads/master.zip) + - Users not familiar with [git](https://git-scm.com/downloads) and GitHub can simply download the repository as a zip file (). In either case, the download is very large (over 900MB compressed download and 3.4GB uncompressed on disk). 3. Start VisionEval by double clicking the `VisionEval.RProj` file in the newly cloned `VisionEval-Source` directory.