Improve exception handling

[chriscarrollsmith]

Investigate and implement different and more appropriate exception types, as we're currently handling too many different cases with 'ValueError'.

Also, it looks like when the API returns HTML instead of JSON, we're not always catching it, as in this error case from imf_databases:

---

KeyError Traceback (most recent call last)
Cell In[3], line 9
7 databases = pd.read_parquet(cache_path)
8 else:
----> 9 databases = imfp.imf_databases()
10 os.makedirs("data", exist_ok=True)
11 databases.to_parquet(cache_path)

File ~\Software\Python\imfp\imfp\data.py:37, in imf_databases(times)
32 url = "http://dataservices.imf.org/REST/SDMX_JSON.svc/Dataflow"
33 raw_dl = _download_parse(url, times)
35 database_id = [
36 dataflow["KeyFamilyRef"]["KeyFamilyID"]
---> 37 for dataflow in raw_dl["Structure"]["Dataflows"]["Dataflow"]
38 ]
40 description = [
41 dataflow["Name"]["#text"]
42 for dataflow in raw_dl["Structure"]["Dataflows"]["Dataflow"]
43 ]
44 database_list = DataFrame({"database_id": database_id, "description": description})

KeyError: 'Structure'