Analysis table fields 'analysisresults__name' and 'analysisresults__value' not filtering properly

[enmust]

@ollehints @urtson This is the same problem like Preparation table has with 'taxonlist' fields.
This link should only show Analysed components 'Zr' which have contents over 100.
http://arendus.geokogud.info/analysis?sample_number_1=5&sample_number=&locality_1=1&locality=&stratigraphy_1=1&stratigraphy=&stratigraphy_bed_1=1&stratigraphy_bed=&analysis_1=1&analysis=&component_1=1&component=Zr&sample_id_1=1&sample_id=&country_1=1&country=&content_1=7&content=100&dbs%5B%5D=1&dbs%5B%5D=2&dbs%5B%5D=3&dbs%5B%5D=4&dbs%5B%5D=5&dbs%5B%5D=6&currentTable=analysis&maxSize=5&sort=id&sortdir=DESC

Here are API links, which show differences
https://api.arendus.geokogud.info/analysis/?paginate_by=30&page=1&order_by=-id&distinct=true&analysisresults__name__icontains=Zr&analysisresults__value__gt=100

https://api.arendus.geokogud.info/analysis_results/?paginate_by=30&page=1&order_by=-id&distinct=true&name__icontains=Zr&value__gt=100

[urtson]

Currently it is not possible to get desired results directly from 'analysis' table as it is an OR search from related table. It will search for analyses where zirconium is measured and analyses where any measured value is above 100.

Possible workaround is to use two queries:

1. https://api.arendus.geokogud.info/analysis_results/?name__icontains=Zr&value__gt=100&fields=id
2. https://api.arendus.geokogud.info/analysis/?analysisresults__id__in=1787,2228,...etc.
   If you get a single record from first query, use without __in lookup

Have to test it for speed, maybe not a good idea. The length of id list will be in thousands.

No idea yet how to do it in one query. Nested queries can be used in Django but it will make two queries anyway.
Another possibility is to create additional analysis summary table with all analyses and results inside.

All above goes also for #62

Also note:
https://api.arendus.geokogud.info/analysis_results/?paginate_by=30&page=1&order_by=-id&distinct=true&name__icontains=Zr&value__gt=100
it's not correct to use distinct=true here as it will eliminate possible distinct records with similar value