20191008 | 20200603
This is short documentation describing how to use R programming language to create an EML document from an Excel metadata workbook (".xlsx") for the ecological data archive. The "R" script reads the metadata Excel workbook, and generates an EML file in ".xml" format.
If you are interested in learning the EML generation steps by watching the video, here it is: https://www.youtube.com/watch?v=rn8Uee49LsM&t=1s
This EDI data packet generation script requires these packages installed with the latest version: 1) EMLassemblyline, 2) XML 3) readxl and 4) dplyr.
The datatables need to be well organized and meet the best practice: such as no special character in heading, consistent column format, clear defined "NA" values...
1.EML_generation
All functions stored in EML_funs
2.Metadata - Metadata Excel workbook, and its instruction
3.project.99 and project.100 as examples
EML_generation stores the main "R" functions and template information for the creation of the ".xml" file. The functions in the "R" script and codes navigate among different files based on their relative locations, the name and location of the folder should not be changed (unless you also want to make changes on the file paths for all "R" functions used).
For adding a new dataset in the Excel workbook, you need to fill in the first 7 sheets, from left to right. The 4 sheets at the end are lists of peoples (may need to update once a while if you have new personnel) and for vocabulary controls. Refer to the "Metadata_Readme.md" for a detailed guide on each sheet of the "Metadata.xlsx".
For a new dataset, create its own project folder called "project.XXX" ("XXXX" = dataset ID). Here, the project.99 is a vegetation dislodgment data package as an example.
In the project folder, we have to have "abstract.docx", "methods.docx", the dataset(s) you want to archive, and an "R" script called "project.XXXX.R" ("XXXX" = dataset ID).
Make sure the "abstract.docx" document is the abstract for the data instead of the abstract for article publications. If the dataset has a protocol, include a URL within the methods document.
As of today, 07/26/2020, the abstract and methods document can't handle bold and hyperlink. Make sure to remove those before generating the EML. Otherwise your EML will be invalid.
Once you have all the metadata information filled out and all the relevant files placed in the project folder, run the "project.XXXX.R" in the project folder, from top to bottom to generate your EML document. The script will generate error messages if any part of the validation fails while "R" would also provide additional "warnings()" and additional messages should it arise when executing the functions.
The R code and functions here are built on top of the EMLassemblyline. So if you update the EMLassemblyline package with the major revision, the code could break. See the GitHub repo: https://github.com/EDIorg/EMLassemblyline.
This workflow provide two intellectual rights: CC-By and CC0. If you want to use other data publication license, you want to modify the R function "generate_EML_Assemblyline" below L123 to read in a special license that in a Word document.
If you manage more than 50 datasets, another similar tool I want to recommend is the Core-metabase. This is the tool developed using PostgresSQL which has great benefits from additional controls and tables (vocabulary controls or extra tables for long-term data management purposes). The core-metabase means to manage a large number of datasets and potentially connect to your website as a backend database. The installation code for the LTER-core-metabase is here: https://github.com/lter/LTER-core-metabase. An R package has developed parallel to the LTER-core-metabase for generating EML: https://github.com/BLE-LTER/MetaEgress.
Ecological Metadata Language (EML) standards and specification can be seen: https://knb.ecoinformatics.org/external//emlparser/docs/index.html