Soxspipe

* **FIXED:** fixing string literal bug resulting from new sqlite3 release

* **ENHANCEMENT:** Pipeline name and version added to product headers.
* **ENHANCEMENT:** recipe ID added to product headers.
* **ENHANCEMENT:** added comments to all settings in the SOXS default settings file
* **ENHANCEMENT:** there are now 2 rounds of object trace detection. The first round helps to locate the slit-position and a typical standard-deviation of the object profile, and the second round uses this information to preform a better measurement and fit of the trace. This is improving the fitting for faint objects.
* **REFACTOR:** adjusted default 'slice' lengths used to detect a continuum in nodding frames.
* **REFACTOR:** many adjustments to improve robustness and extractions
* **FIXED:** recipe command was getting written to the bottom of the log file of the previous recipe instead of the top of the current recipe log

* **ENHANCEMENT:** resolution plots added to the `soxs-spatial-solution` QC PDF (if a through-slit arc frame file is provided).

* **ENHANCEMENT:** Major updates to the documentation following PAE.

* **ENHANCEMENT:** Raw frames & their categories are reported in the FITS headers of the recipe products.
* **ENHANCEMENT:** Intermediate calibration products are reported in the recipe product FITS headers alongside their MD5 hash (DATAMD5)
* **ENHANCEMENT:** Recipe settings and values are reported in the FITS headers of the recipe products.
* **ENHANCEMENT:** The recipe command is now printed before AND after execution to the terminal and the recipe log file.
* **REFACTOR:** The length of SOF and product filenames have been compressed so they can be reported within the FITS header character limit.
* **REFACTOR:** reduced products are now stored in the ESO-compliant nested-folder scheme `/reduced/YYYY-MM-DD/`
* **REFACTOR:** QCs are also stored in a similar nested-folder scheme `/qc/YYYY-MM-DD/`
* **FIXED:** All FITS file products pass fitsverify checks

* **REFACTOR:** updating the data-organiser to fix a performance issue.
* **FIXED:** Allow the object trace polynomial orders to dynamically change if a fit to the continuum is not found (ala the order-centre tracing)