nv-ingest error

[ashish-kumar-hpe]

Sample Code

from nv_ingest_client.client import Ingestor

ingestor = (
Ingestor(message_client_hostname="192.168.21.179",message_client_port=30415)
.files("./data/a00142461enw.pdf")
.extract(
extract_text=True,
extract_tables=False,
extract_images=False,
)
.embed()
.vdb_upload()
)

results = ingestor.ingest()
print(results)

Error in nv-ingest

ERROR: Error: 1 validation error for IngestJobSchema
tasks -> 1 -> root
'NoneType' object is not callable (type=type_error)
Traceback (most recent call last):
File "/opt/conda/envs/nv_ingest/lib/python3.10/site-packages/nv_ingest/api/v1/ingest.py", line 138, in submit_job
submitted_job_id = await ingest_service.submit_job(updated_job_spec)
File "/opt/conda/envs/nv_ingest/lib/python3.10/site-packages/nv_ingest/service/impl/ingest/redis_ingest_service.py", line 60, in submit_job
validate_ingest_job(job_spec)
File "/opt/conda/envs/nv_ingest/lib/python3.10/site-packages/nv_ingest/schemas/ingest_job_schema.py", line 204, in validate_ingest_job
return IngestJobSchema(**job_data)
File "pydantic/main.py", line 341, in pydantic.main.BaseModel.init
pydantic.error_wrappers.ValidationError: 1 validation error for IngestJobSchema
tasks -> 1 -> root
'NoneType' object is not callable (type=type_error)

[Advait251206]

This error is coming from Pydantic schema validation inside nv-ingest, not from your PDF or embedding logic:

ValidationError: IngestJobSchema
tasks -> 1 -> root
'NoneType' object is not callable

That specific message almost always means:

❗ One of the pipeline task objects is malformed or missing a required callable field.

---

1. What is actually breaking

nv-ingest builds an internal pipeline like:

files → extract → embed → vdb_upload

Each step becomes a task object with a “root callable” function.

Your error:

tasks -> 1 -> root is None

means:

👉 The second task in the pipeline is missing its handler function

---

2. Most likely cause in your code

Your chain:

Ingestor(...)
.files(...)
.extract(...)
.embed()
.vdb_upload()

looks correct, BUT in nv-ingest:

❗ .embed() and .vdb_upload() require proper configuration context

If ANY of the following are missing:

• embedding endpoint config
• vector DB config
• default pipeline registry

then nv-ingest may create a None task internally, which later crashes validation.

---

3. Common root causes

✔ (A) Missing embedding config

You did not specify:

• endpoint_url
• model_name

So .embed() may be creating a null embedding task

---

✔ (B) Missing VDB configuration

.vdb_upload() requires:

• vector DB backend config (Milvus / FAISS / Elastic / etc.)
• index name / connection config

Without it → task is created as None

---

✔ (C) Version mismatch bug (very common in nv-ingest)

This error has been seen when:

• nv_ingest_client version ≠ nv_ingest server version
• schema expects different task structure

Result:

schema builds task list incorrectly → root becomes None

---

4. Why the error looks confusing

This part:

'NoneType' object is not callable

is misleading.

It is NOT your code calling None.

It means:

A task's "root function" field was expected to be callable, but is None during schema validation.

---

5. How to fix it

✔ Fix 1: explicitly configure embed()

Try:

.embed(
    endpoint_url="http://<your-embed-service>:9997/v1",
    model_name="bge-m3"
)

Without this, embed task may be invalid.

---

✔ Fix 2: explicitly configure VDB

Example (depends on backend):

.vdb_upload(
    vdb_type="milvus",
    collection_name="test_collection",
    host="localhost",
    port=19530
)

If omitted → task may become None.

---

✔ Fix 3: verify version compatibility

Run:

pip show nv-ingest-client
pip show nv-ingest

Make sure:

client and server versions match exactly

---

✔ Fix 4: isolate the failing stage

Test step-by-step:

.files(...)
.extract(...)

Then:

.embed()

Then:

.vdb_upload()

This will identify which stage creates the invalid task.

---

6. Most likely root cause (based on your snippet)

Your pipeline:

.embed()
.vdb_upload()

without parameters strongly suggests:

❗ missing embedding + vector DB configuration → task becomes None → schema crash

---

7. Summary

• ❌ Not a PDF issue
• ❌ Not ingestion failure
• ❗ Pipeline task is malformed (None root callable)
• ✔ Usually caused by missing embed/VDB config or version mismatch

---

If you want

I can rewrite your full pipeline into a known-good nv-ingest configuration for:

• local embedding (bge-m3)
• FAISS or Milvus backend
• working end-to-end ingestion

---

If this answer helped or pointed you in the right direction, I'd appreciate it if you could mark it as the accepted answer so it's easier for others with the same issue to find.

Also, if you found my contribution useful, I'd appreciate it if you could check out my GitHub profile, follow me, and star any repositories you find interesting.

GitHub: https://github.com/Advait251206