add suport for image build

[alexcos20]

Fixes #1026

Changes proposed in this PR:

1. Allow c2d algorithm image to be build on the fly, by adding dockerfile parameter to algorithm.meta.container

You can also add additional files, using additionalDockerFiles, which has file name as a key and file content

TBD: Potential Memory Issue
If additionalDockerFiles contains large files, all are loaded into memory at once. For large files, consider streaming them into the tar archive.

Example:

{
    "command": "freeStartCompute",
    "datasets":[],
    "algorithm": {
        "meta": {
            "container": {
                "dockerfile":"FROM rocm/dev-ubuntu-24.04 as base\r\nRUN apt-get update && apt-get -y install bash curl git wget python3 build-essential\r\n\r\nCOPY . /benchmark\r\nWORKDIR /benchmark\r\nRUN pip install -r requirments.txt --break-system-packages",
                "additionalDockerFiles":{
                    "requirments.txt":"pyopencl>=2023.1\r\nnumpy>=1.24"
                }
                
            }
        },
        "envs": {
            "SEED":"test"
        }

    },
    "consumerAddress": "0x00",
    "signature": "123",
    "nonce": 1,
    "environment": "0x956cb02fee19ac8466aea8cf1901e3b771b906ed7bc1fb868c3854de6fa338e5-0x14c61b597679187ca041a8eee95d0d5167e550d313a90c29c3867997f3e4ee1c",
    "resources": [
        {
            "id": "cpu",
            "amount": 1
        },
        {
            "id":"disk",
            "amount":10000
        }
    ]
}

2. Add a new type of result: imageLog , when executing getComputeStatus:

 {
        "owner": "0x00",
        "did": null,
        "jobId": "2a66f0bf4587bbb7f2824f6feedc0137240ff2fbd8166e813dcac51fa772791d",
        "dateCreated": "1757421862.13",
        "dateFinished": null,
        "status": 40,
        "statusText": "Running algorithm ",
        "results": [
            {
                "filename": "image.log",
                "filesize": 199,
                "type": "imageLog",
                "index": 0
            }
        ]
}

Sweet: you can always call getResults and pass image.log index, to get a realtime log from image build or pull process, which is especially handy when building large images using dockerfile option

3. changes in processLoop

Removed image pulling from there, as it's time consuming and blocks entire loop.
Image pull/build is now handled in startCompute, before job is handled by main loop

[alexcos20]

Now, we can add files to the building process.
Query example:

{
    "command": "freeStartCompute",
    "datasets":[],
    "algorithm": {
        "meta": {
            "container": {
                "image": "oceanprotocol/node-benchmark",
                "tag":"latest",
                "dockerfile":"FROM rocm/dev-ubuntu-24.04 as base\r\nRUN apt-get update && apt-get -y install bash curl git wget python3 build-essential\r\n\r\nCOPY . /benchmark\r\nWORKDIR /benchmark\r\nRUN pip install -r requirments.txt --break-system-packages",
                "additionalDockerFiles":{
                    "requirments.txt":"pyopencl>=2023.1\r\nnumpy>=1.24"
                }
                
            }
        },
        "envs": {
            "SEED":"test"
        }

    },
    "consumerAddress": "0x123",
    "signature": "123",
    "nonce": 1,
    "environment": "0x956cb02fee19ac8466aea8cf1901e3b771b906ed7bc1fb868c3854de6fa338e5-0x5d45d99c3964a3141fa4463f6381697f452edbd6fb1dcce866ce602e0bb43477",
    "resources": [
        {
            "id": "cpu",
            "amount": 1
        },
        {
            "id":"disk",
            "amount":10000
        }
    ]
}

Result:

[alexcos20]

Ocean.js update: oceanprotocol/ocean.js#1993

[giurgiur99]

LGTM!