this piece of code returns records from zenodo with no problem

import requests
import json

records_api_url = 'https://zenodo.org/api/records'
search_query = 'grants.acronym:DIGITAF AND resource_type.type:publication'
res = requests.get(records_api_url, params={'q': search_query, 'size': 10000})
for i in res.json()["hits"]["hits"]:
    print(i["metadata"], "n" )

returning about 20 hits (dictionaries), one of them, as example here:

{
    "title": "Agroforestry & the Framework for Certification of Carbon Removals",
    "doi": "10.5281/zenodo.7828353",
    "publication_date": "2022-12-31",
    "description": "<p>Policy Briefing #20 welcomes the draft CRCF, and its initial proposals for robust certification of carbon removals including a) quality criteria, b) verification and certification, c) functioning of certification schemes.  However we feel that rigorous certification is only needed for a subset of specialised
 carbon farming schemes.  There is much more to be done, in the near term, to meet the aspirational target of the December 2021 “Sustainable Carbon Cycles Communication” that “by 2028 every land manager should have access to verified emission and removal data”.  EURAF believes that high quality emissions estimates by Member States (IPCC Tier 3) should be combined with detailed wall-to-wall representation of land use (IPCC Approach 3), and results should be available open source to farmers and foresters.  High-resolution databases of net emissions across the forestry and agricultural sectors will facilitate a target of net zero emissions in the integrated land sector by 2035. These improvements by MS should proceed in parallel with amendment and ratification of the CRCF.</p>",
    "access_right": "open",
    "creators": [
        {
            "name": "LAWSON, Gerry",
            "affiliation": "EURAF",
            "orcid": "0000-0002-1395-3092"
        }
    ],
    "keywords": [
        "carbon farming",
        "agroforestry",
        "DigitAF"
    ],
    "version": "1",
    "resource_type": {
        "title": "Working paper",
        "type": "publication",
        "subtype": "workingpaper"
    },
    "license": {
        "id": "cc-by-4.0"
    },
    "grants": [
        {
            "code": "101059794",
            "internal_id": "10.13039/501100000780: : 101059794",
            "funder": {
                "name": "European Commission",
                "doi": "10.13039/501100000780",
                "acronym": "EC"
            },
            "title": "DIGItal Tools to help AgroForestry meet climate, biodiversity and farming sustainability goals: linking field and cloud",
            "acronym": "DIGITAF",
            "program": "HE",
            "url": "https: //cordis.europa.eu/projects/101059794"
        }
    ],
    "communities": [
        {
            "id": "euraf"
        },
        {
            "id": "eu"
        }
    ],
    "relations": {
        "version": [
            {
                "index": 0,
                "is_last": True,
                "parent": {
                    "pid_type": "recid",
                    "pid_value": "7828352"
                }
            }
        ]
    },
    "notes": "EURAF Policy Briefing #20"
}

I have no problem accessing the title from this dict with

print(i["metadata"]["title"], "n" )

However, when trying accessing the description with

print(i["metadata"]["description"], "n" )

returns a “KeyError: ‘description'” which is puzzling me. What am I missing?

print(i["metadata"]["title"] # works fine
print(i["metadata"]["doi"] # works fine
print(i["metadata"]["publication_date"] # works fine

print(i["metadata"]["description"] # DOES NOT WORK -> triggers a KeyError: 'description'