fkr / dayone2logseq-md

Converting DayOne journals into markdown files compatible with Logseq

dayone2markdown (logseq-optimized version)

Converting DayOne journals (from JSON file) into Markdown files. Based on original version from @ploum. Compatible with Python3. The script is intended for conversion of journals into a plain-text format compatible with Logseq (md syntax).

Usage

1. Export you DayOne journal as JSON and unzip the folder.
2. Put the script in the unzipped folder.
3. Run the script python do2md.py Journal.json
4. Output .md files are created in "do2md" subfolder. You can either put the files directly to /journals, or keep them in the sub-folder for better organization. Rename the folder with photos to "dayone" and move it into Logseq /assets folder. Re-index your LS database, and the files should now be in your graph.

Results :

1. Each entry is now converted to an MD file. Name of the file is the date-time of the entry compatible with Logseq's yyyy_mm_dd.md expected format. IF there are multiple entries at the same day, they are appended into a single file.
2. Pictures are inserted with a relative path compatible with Logseq ( assets/dayone/)
3. [not tested in this version, inherited from original version] Tags are inserted in the text as " #tag" (if they were not previously)
4. Physical location of an entry (if present) is written into collapsed block at the end of the file with both address and coordinates.
5. The script attemps to standardize entry headers into somewhat uniform format: if the entry beggins with a header, it is preserved. If not, it attempts to find a linebreak in first 30 characters in the entry, and use the first paragraph as header. If that is not the case, a "DayOne Entry" header is used. A hashtag (#diary) is appended at the beginning of every header, to backlink all the journal entries from the DayOne journal together.
6. The full time of the entry is saved as Logseq page property (publication-date::).

Limitations

1. I attempted to the fix issues occuring with formatting (multiple newlines etc), but formatting issues may still persist. I'll try to fix those over time. 1b. The script is intended to put each paragraph in the entry as new child-block, to make the reading experience in Logseq more streamlined. However, this does not seem to work 100% of time. Will attempt to fix it once I figure out why.
2. Metadata other than tags, time and location are lost.

Work in progress.