Make GCN draft even if info is missing

[JannisNe]

Make draft_gcn() run even if info like the observation log is missing and raise warnings.
For IC230112A for example, nu.plot_overlap_with_observations(first_det_window_days=3.) will give

INFO:nuztf.observations:Getting observation logs from skyvision.
/Users/jannisnecker/opt/miniconda3/envs/nuztf/lib/python3.10/site-packages/ztfquery/skyvision.py:240: UserWarning: No observing log for 2023-01-12
  warnings.warn(f"No observing log for {date}")
/Users/jannisnecker/opt/miniconda3/envs/nuztf/lib/python3.10/site-packages/ztfquery/skyvision.py:240: UserWarning: No observing log for 2023-01-13
  warnings.warn(f"No observing log for {date}")
/Users/jannisnecker/opt/miniconda3/envs/nuztf/lib/python3.10/site-packages/ztfquery/skyvision.py:240: UserWarning: No observing log for 2023-01-14
  warnings.warn(f"No observing log for {date}")
/Users/jannisnecker/opt/miniconda3/envs/nuztf/lib/python3.10/site-packages/ztfquery/skyvision.py:240: UserWarning: No observing log for 2023-01-15
  warnings.warn(f"No observing log for {date}")
WARNING:nuztf.base_scanner:No observations found for this event.
WARNING:nuztf.base_scanner:Not plotting overlap with observations.

and nu.draft_gcn()

WARNING:nuztf.base_scanner:No overlap line added!

with the resulting draft

Astronomer Name (Institute of Somewhere), ............. report,

On behalf of the Zwicky Transient Facility (ZTF) and Global Relay of Observatories Watching Transients Happen (GROWTH) collaborations: 

As part of the ZTF neutrino follow up program (Stein et al. 2022), we observed the localization region of the neutrino event IceCube-230112A (Massimiliano et. al, GCN 33161) with the Palomar 48-inch telescope, equipped with the 47 square degree ZTF camera (Bellm et al. 2019, Graham et al. 2019). We started observations in the g- and r-band beginning at 2023-01-13 02:43 UTC, approximately 20.0 hours after event time. Each exposure was 300s with a typical depth of 21.0 mag. 
 
The images were processed in real-time through the ZTF reduction and image subtraction pipelines at IPAC to search for potential counterparts (Masci et al. 2019). AMPEL (Nordin et al. 2019, Stein et al. 2021) was used to search the alerts database for candidates. We reject stellar sources (Tachibana and Miller 2018) and moving objects, and apply machine learning algorithms (Mahabal et al. 2019) . 

No candidate counterparts were detected. 

ZTF and GROWTH are worldwide collaborations comprising Caltech, USA; IPAC, USA; WIS, Israel; OKC, Sweden; JSI/UMd, USA; DESY, Germany; TANGO, Taiwan; UW Milwaukee, USA; LANL, USA; TCD, Ireland; IN2P3, France.

GROWTH acknowledges generous support of the NSF under PIRE Grant No 1545949.
Alert distribution service provided by DIRAC@UW (Patterson et al. 2019).
Alert database searches are done by AMPEL (Nordin et al. 2019).
Alert filtering is performed with the nuztf (Stein et al. 2021, https://github.com/desy-multimessenger/nuztf).

[coveralls]

Coverage: 82.429% (-0.2%) from 82.646% when pulling 300ea18 on make_gcn_anyway into 8aa7afe on main.