### Install openreview-downloader

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Install the package using pip. This is the first step before using the CLI.

```bash
pip install openreview_downloader
```

--------------------------------

### Install openreview-downloader

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Install the tool using pip. For development, install from source with development dependencies.

```bash
pip install openreview_downloader
```

```bash
pip install -e '.[dev]'
```

--------------------------------

### Install development dependencies and run tests

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Installs the package in editable mode with development dependencies and discovers and runs unit tests.

```bash
pip install -e '.[dev]'
python -m unittest discover -s tests

# Expected output:
# ....
# ----------------------------------------------------------------------
# Ran 4 tests in 0.003s
# OK
```

--------------------------------

### Install Development Dependencies

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Install the project in editable mode with all development dependencies included. This is necessary for contributing to the project.

```bash
pip install -e '.[dev]'
```

--------------------------------

### CLI Usage Examples

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Examples of using the `ordl` command-line tool to download accepted papers with specific criteria, including regex matching, case sensitivity, and JSONL output format.

```APIDOC
## CLI Usage Examples

### Require both a regex AND a text term
```bash
ordl accepted --list --regex 'protein(s)?' --grep graph \
     --venue-id NeurIPS.cc/2025/Conference
```

### Case-sensitive regex download
```bash
ordl accepted --regex 'GPT-[0-9]+' --case-sensitive \
     --venue-id NeurIPS.cc/2025/Conference
```

### `--format jsonl` — Machine-readable JSONL output
When `--list` is active, `--format jsonl` emits a summary line followed by one JSON object per paper on stdout. Progress and status messages go to stderr so stdout can be piped cleanly.

```bash
ordl accepted --list --search diffusion --head 2 \
     --format jsonl --venue-id NeurIPS.cc/2025/Conference \
     > papers.jsonl

# stdout (papers.jsonl):
# {"decisions": ["accepted"], "head": 2, "matched_papers": 710,
#  "shown_papers": 2, "type": "summary", "venue_id": "NeurIPS.cc/2025/Conference"}
# {"authors": "Jiyao Zhang, ...", "decision": "accepted", "id": "CB8jwNE2vV",
#  "match_count": 1,
#  "matches": [{"count": 1, "field": "abstract", "query": "diffusion",
#               "snippet": "...occupancy-[diffusion] model..."}],
#  "number": 29119,
#  "pdf_path": "downloads/neurips2025/accepted/29119_CADGrasp_...pdf",
#  "title": "CADGrasp: ...", "type": "paper",
#  "venue": "NeurIPS 2025 poster", "venueid": "NeurIPS.cc/2025/Conference"}

# Pipe into jq for lightweight post-processing
ordl accepted --list --format jsonl --venue-id ICLR.cc/2025/Conference \
     2>/dev/null | jq 'select(.type=="paper") | .title'
```

### Other conferences

The `--venue-id` flag accepts any OpenReview venue identifier.

```bash
# ICLR 2025 orals only
ordl oral --venue-id ICLR.cc/2025/Conference

# ICML 2025 oral + spotlight
ordl oral,spotlight --venue-id ICML.cc/2025/Conference

# Info for ICLR
ordl --info --venue-id ICLR.cc/2025/Conference
```
```

--------------------------------

### JSON Lines Output Example

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Example of JSON Lines output, which includes a summary line followed by individual paper records. Each paper record contains stable fields for easy parsing.

```json
{"decisions": ["accepted"], "head": 2, "matched_papers": 710, "shown_papers": 2, "type": "summary", "venue_id": "NeurIPS.cc/2025/Conference"}
{"authors": "Jiyao Zhang, Zhiyuan Ma, Tianhao Wu, Zeyuan Chen, Hao Dong", "decision": "accepted", "id": "CB8jwNE2vV", "match_count": 1, "matches": [{"count": 1, "field": "abstract", "query": "diffusion", "snippet": "... high-dimensional representation, we introduce an occupancy-[diffusion] model with voxel-level conditional guidance and force closu..."}], "number": 29119, "pdf_path": "downloads/neurips2025/accepted/29119_CADGrasp_Learning_Contact_and_Collision_Aware_General_Dexterous_Grasping_in_Cluttered_Scenes.pdf", "title": "CADGrasp: Learning Contact and Collision Aware General Dexterous Grasping in Cluttered Scenes", "type": "paper", "venue": "NeurIPS 2025 poster", "venueid": "NeurIPS.cc/2025/Conference"}
```

--------------------------------

### Get Decision Counts Without Downloading

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Fetch and display the counts of papers for each decision category without downloading any PDFs. Useful for getting an overview of submissions.

```bash
ordl --info --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Create an OpenReview client

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Use `build_client()` to create an `OpenReviewClient`. It can be anonymous or authenticated by setting `OPENREVIEW_USERNAME` and `OPENREVIEW_PASSWORD` environment variables.

```python
import os
from openreview_downloader.cli import build_client

# Anonymous client (sufficient for public venues)
client = build_client()

# Authenticated client (set env vars before calling)
os.environ["OPENREVIEW_USERNAME"] = "user@example.com"
os.environ["OPENREVIEW_PASSWORD"] = "s3cr3t"
client = build_client()
```

--------------------------------

### Python API: build_client()

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Creates an authenticated OpenReview client. It can read credentials from environment variables `OPENREVIEW_USERNAME` and `OPENREVIEW_PASSWORD` for authenticated access.

```APIDOC
## Python API

All CLI logic is importable from `openreview_downloader.cli`. Below are the key public functions.

### `build_client()` — Create an authenticated OpenReview client

Returns an `openreview.api.OpenReviewClient` pointed at `api2.openreview.net`. Reads optional credentials from environment variables `OPENREVIEW_USERNAME` and `OPENREVIEW_PASSWORD`.

```python
import os
from openreview_downloader.cli import build_client

# Anonymous client (sufficient for public venues)
client = build_client()

# Authenticated client (set env vars before calling)
os.environ["OPENREVIEW_USERNAME"] = "user@example.com"
os.environ["OPENREVIEW_PASSWORD"] = "s3cr3t"
client = build_client()
```
```

--------------------------------

### Run Unit Tests

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Execute the project's unit tests using the unittest module. Ensure all tests pass to verify code integrity.

```bash
python -m unittest discover -s tests
```

--------------------------------

### List First N Papers (Default to Accepted)

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

If no decision is specified when using the --list flag, the CLI defaults to listing accepted papers. This command shows the first 20 accepted papers.

```bash
ordl --head 20 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Search Accepted Papers by Regex

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Preview accepted papers matching a regex pattern. This allows for more complex matching criteria than simple text search. Use `--list` to preview without downloading.

```bash
ordl accepted --list --regex 'diffusion|transformer' --head 2 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Download Papers Matching Text Search

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download papers that match a text search query. This command reruns the same query as the preview command but without the `--list` flag, initiating the download.

```bash
ordl accepted --search diffusion --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Download Limited Number of Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download only the first N matching papers for a given search query. Useful for testing or downloading a small subset.

```bash
ordl accepted --search diffusion --head 10 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Python API: fetch_notes()

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Fetches raw submission notes (accepted and rejected) from the OpenReview API for a given venue.

```APIDOC
### `fetch_notes(client, venue_id, need_rejected)` — Fetch raw submission notes

Returns `(accepted_notes, rejected_notes)` lists from the OpenReview API.

```python
from openreview_downloader.cli import build_client, fetch_notes

client = build_client()
venue_id = "NeurIPS.cc/2025/Conference"

accepted, rejected = fetch_notes(client, venue_id, need_rejected=True)
print(f"Accepted: {len(accepted)}, Rejected: {len(rejected)}")
# Accepted: 5286, Rejected: 254
```
```

--------------------------------

### split_existing(selected, skip_existing)

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Separates a list of selected items into those that need to be downloaded and those that already exist. If `skip_existing` is True, items already present on disk are excluded from the download queue.

```APIDOC
## split_existing(selected, skip_existing)

### Description
Separate new vs. already-downloaded files. Returns `(to_download, existing_count)`. When `skip_existing=True`, files already on disk are omitted from the download queue.

### Parameters
- **selected** (list): A list of selected items, typically from `collect_selected`.
- **skip_existing** (bool): If True, files that already exist on disk will be excluded from the `to_download` list.

### Returns
A tuple containing:
- **to_download** (list): A list of items that need to be downloaded.
- **existing_count** (int): The number of items that already exist on disk.
```

--------------------------------

### Download papers from specific conferences and decisions

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Use the `--venue-id` flag to specify conferences and comma-separated values for decisions like `oral` or `spotlight`.

```bash
# ICLR 2025 orals only
ordl oral --venue-id ICLR.cc/2025/Conference

# ICML 2025 oral + spotlight
ordl oral,spotlight --venue-id ICML.cc/2025/Conference

# Info for ICLR
ordl --info --venue-id ICLR.cc/2025/Conference
```

--------------------------------

### Download accepted papers with case-sensitive regex

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Perform a case-sensitive regex search for accepted papers using the `--regex` and `--case-sensitive` flags. Specify the venue with `--venue-id`.

```bash
ordl accepted --regex 'GPT-[0-9]+' --case-sensitive \
     --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Search Accepted Papers by Text

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Preview accepted papers matching a text query. Use `--list` to preview without downloading. The `--head` option limits the number of results shown.

```bash
ordl accepted --list --search diffusion --head 2 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Derive output directory from venue ID

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Generates a standardized directory path for downloads based on the conference venue ID.

```python
from openreview_downloader.cli import conference_dir

print(conference_dir("NeurIPS.cc/2025/Conference"))
# downloads/neurips2025

print(conference_dir("ICLR.cc/2024/Conference"))
# downloads/iclr2024
```

--------------------------------

### List Accepted and Rejected Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

List both accepted and rejected papers for a venue without downloading. This provides a comprehensive list of all submissions.

```bash
ordl all --list --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### List First N Accepted Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

List only the first specified number of accepted papers. This is useful for quickly previewing a small subset of the results.

```bash
ordl accepted --list --head 3 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### List papers without downloading

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Prints a formatted list of matching papers to standard output. Can be combined with filters like --head, --search, and --regex. Omitting DECISIONS defaults to 'accepted' and enables --list.

```bash
# List first 3 accepted papers
ordl accepted --list --head 3 --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Omit DECISIONS to default to accepted and auto-enable --list
ordl --head 20 --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Download NeurIPS Oral and Spotlight Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download papers from multiple decision categories (oral and spotlight) for a given venue.

```bash
ordl oral,spotlight --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Build a download list for selected papers

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Use `collect_selected` to generate a list of papers matching specific decisions. It returns tuples containing the note, category, and desired download path, deduplicating by note ID.

```python
from pathlib import Path
from openreview_downloader.cli import (
    build_client, fetch_notes, collect_selected
)

client = build_client()
venue_id = "NeurIPS.cc/2025/Conference"
accepted, rejected = fetch_notes(client, venue_id, need_rejected=False)

selected = collect_selected(
    accepted=accepted,
    rejected=rejected,
    venue_id=venue_id,
    decisions=["oral", "spotlight"],
    base_dir=Path("downloads/neurips2025"),
)
for note, category, path in selected[:3]:
    print(f"[{category}] {path.name}")
# [oral] 27970_Deep_Compositional_Phase_Diffusion.pdf
# [spotlight] 28003_Some_Spotlight_Paper.pdf
```

--------------------------------

### Download ICML 2025 Oral and Spotlight Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download papers that received either an 'oral' or 'spotlight' decision for ICML 2025. Multiple decisions can be specified as a comma-separated list.

```bash
ordl oral,spotlight --venue-id ICML.cc/2025/Conference
```

--------------------------------

### Fetch raw submission notes

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

The `fetch_notes` function retrieves accepted and rejected paper notes from the OpenReview API for a given venue. Set `need_rejected=True` to also fetch rejected papers.

```python
from openreview_downloader.cli import build_client, fetch_notes

client = build_client()
venue_id = "NeurIPS.cc/2025/Conference"

accepted, rejected = fetch_notes(client, venue_id, need_rejected=True)
print(f"Accepted: {len(accepted)}, Rejected: {len(rejected)}")
# Accepted: 5286, Rejected: 254
```

--------------------------------

### Separate existing files from download queue

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Determines which files need to be downloaded by checking for their existence on disk. Returns the count of files to download and the count of already existing files.

```python
from pathlib import Path
import tempfile
from openreview_downloader.cli import split_existing
from types import SimpleNamespace

with tempfile.TemporaryDirectory() as tmp:
    existing_path = Path(tmp) / "accepted" / "paper.pdf"
    existing_path.parent.mkdir()
    existing_path.write_bytes(b"%PDF-1.4")

    note = SimpleNamespace(id="note-1", number=1, content={})
    selected = [(note, "accepted", existing_path, None)]

    to_download, already_present = split_existing(selected, skip_existing=True)
    print(f"To download: {len(to_download)}, Already present: {already_present}")
    # To download: 0, Already present: 1
```

--------------------------------

### List Accepted Papers Without Downloading

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

List all accepted papers for a venue without downloading the PDF files. This command shows paper details and their intended save path.

```bash
ordl accepted --list --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Filter accepted papers by regex and text term

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Use the `--regex` and `--grep` flags to filter accepted papers based on regular expressions and text terms. Specify the venue using `--venue-id`.

```bash
ordl accepted --list --regex 'protein(s)?' --grep graph \
     --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### JSON Lines Output for Automation

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

For scripts, agents, and crawlbots, use JSON Lines output (`--format jsonl`) while listing papers. This format is convenient for programmatic processing.

```bash
ordl accepted --list --search diffusion --head 2 --format jsonl --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Download papers by decision category

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Downloads PDFs for specified decision categories (e.g., oral, spotlight, accepted). Skips existing files by default. Can customize output directory and force re-fetching.

```bash
# Download all NeurIPS 2025 oral presentations
ordl oral --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download both oral and spotlight papers
ordl oral,spotlight --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download every accepted paper
ordl accepted --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download accepted + rejected (shorthand "all")
ordl all --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download to a custom directory, always re-fetching existing files
ordl accepted --venue-id ICLR.cc/2025/Conference \
     --out-dir /data/iclr2025 \
     --no-skip-existing
```

--------------------------------

### Download ICLR 2025 Orals

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download papers with the 'oral' decision for a specific venue. Change the `--venue-id` to target different conferences.

```bash
ordl oral --venue-id ICLR.cc/2025/Conference
```

--------------------------------

### Download All Accepted NeurIPS Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download all papers that were accepted, regardless of their presentation type.

```bash
ordl accepted --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Download papers in JSONL format

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Use `--format jsonl` to output paper data in JSON Lines format. This is useful for machine-readable processing. Progress messages are sent to stderr.

```bash
ordl accepted --list --search diffusion --head 2 \
     --format jsonl --venue-id NeurIPS.cc/2025/Conference \
     > papers.jsonl
```

--------------------------------

### Download All ICLR 2025 Accepted Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download all papers that were accepted for ICLR 2025, regardless of their presentation format. This command fetches all papers with the 'accepted' decision.

```bash
ordl accepted --venue-id ICLR.cc/2025/Conference
```

--------------------------------

### Download NeurIPS Oral Papers

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Download all oral presentation papers for a specific venue. The files are saved in a structured directory.

```bash
ordl oral --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Compile regex patterns for filtering

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Compiles a list of string patterns into regex objects. Raises an error for invalid patterns. Case sensitivity can be controlled.

```python
from openreview_downloader.cli import compile_regexes

patterns = compile_regexes(
    ["diffusion|transformer", r"GPT-\d+"],
    case_sensitive=False,
)
# [re.compile('diffusion|transformer', re.IGNORECASE),
#  re.compile('GPT-\d+', re.IGNORECASE)]
print(patterns)
```

--------------------------------

### conference_dir(venue_id)

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Derives a standardized output directory path from a given venue ID. This function helps in organizing downloaded papers by creating a consistent directory structure.

```APIDOC
## conference_dir(venue_id)

### Description
Derive output directory from venue ID. Returns a `Path` like `downloads/neurips2025` derived from the venue ID.

### Parameters
- **venue_id** (str): The venue identifier string (e.g., "NeurIPS.cc/2025/Conference").

### Returns
A `Path` object representing the output directory for the conference.
```

--------------------------------

### Text search across paper metadata

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Searches papers case-insensitively over title, authors, abstract, keywords, decision, venue, id, and paper number. Repeat the flag for AND logic. Use --grep as an alias.

```bash
# List accepted papers mentioning "diffusion" (preview only)
ordl accepted --list --search diffusion --head 2 \
     --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Require both "diffusion" AND "protein" to match
ordl accepted --list --grep diffusion --grep protein \
     --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download all matches (remove --list)
ordl accepted --search diffusion \
     --venue-id NeurIPS.cc/2025/Conference
```

```bash
# Download only the first 10 matches
ordl accepted --search diffusion --head 10 \
     --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Process JSONL output with jq

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Pipe the JSONL output to `jq` for filtering and extracting specific fields, such as paper titles. Redirect stderr to `/dev/null` to clean up stdout.

```bash
# stdout (papers.jsonl):
# {"decisions": ["accepted"], "head": 2, "matched_papers": 710,
#  "shown_papers": 2, "type": "summary", "venue_id": "NeurIPS.cc/2025/Conference"}
# {"authors": "Jiyao Zhang, ...", "decision": "accepted", "id": "CB8jwNE2vV",
#  "match_count": 1,
#  "matches": [{"count": 1, "field": "abstract", "query": "diffusion",
#               "snippet": "...occupancy-[diffusion] model..."}],
#  "number": 29119,
#  "pdf_path": "downloads/neurips2025/accepted/29119_CADGrasp_...pdf",
#  "title": "CADGrasp: ...", "type": "paper",
#  "venue": "NeurIPS 2025 poster", "venueid": "NeurIPS.cc/2025/Conference"}

# Pipe into jq for lightweight post-processing
ordl accepted --list --format jsonl --venue-id ICLR.cc/2025/Conference \
     2>/dev/null | jq 'select(.type=="paper") | .title'
```

--------------------------------

### Filter selected papers with search/regex criteria

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Applies search and regex filters to collected papers. Ensure 'args' namespace includes 'search_terms' and 'regexes'.

```python
import argparse
from pathlib import Path
from openreview_downloader.cli import (
    build_client, fetch_notes, collect_selected,
    filter_selected, compile_regexes
)

client = build_client()
venue_id = "NeurIPS.cc/2025/Conference"
accepted, _ = fetch_notes(client, venue_id, need_rejected=False)
selected = collect_selected(accepted, [], venue_id,
                            ["accepted"],
                            Path("downloads/neurips2025"))

args = argparse.Namespace(
    venue_id=venue_id,
    decisions=["accepted"],
    search_terms=["diffusion"],
    regexes=compile_regexes(["protein(s)?"], case_sensitive=False),
    case_sensitive=False,
    head=None,
    format="text",
    list=True,
)
matched = filter_selected(selected, args)
print(f"Papers matching 'diffusion' AND /protein(s)?/: {len(matched)}")
for note, cat, path, info in matched[:2]:
    print(f"  {note.number} hits={info['hit_count']} – {path.name}")
```

--------------------------------

### Python API: collect_selected()

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Builds a list of papers matching specific decisions, returning a tuple of the note, its category, and the intended download path.

```APIDOC
### `collect_selected(accepted, rejected, venue_id, decisions, base_dir)` — Build a download list

Returns `[(note, category, path), ...]` — one entry per paper matching the requested decisions. Deduplicates by note ID.

```python
from pathlib import Path
from openreview_downloader.cli import (
    build_client, fetch_notes, collect_selected
)

client = build_client()
venue_id = "NeurIPS.cc/2025/Conference"
accepted, rejected = fetch_notes(client, venue_id, need_rejected=False)

selected = collect_selected(
    accepted=accepted,
    rejected=rejected,
    venue_id=venue_id,
    decisions=["oral", "spotlight"],
    base_dir=Path("downloads/neurips2025"),
)
for note, category, path in selected[:3]:
    print(f"[{category}] {path.name}")
# [oral] 27970_Deep_Compositional_Phase_Diffusion.pdf
# [spotlight] 28003_Some_Spotlight_Paper.pdf
```
```

--------------------------------

### Regular-expression search for papers

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Searches paper metadata using Python regex syntax. Repeat the flag for multiple required patterns. Use --case-sensitive for exact matching.

```bash
# List papers matching either "diffusion" or "transformer"
ordl accepted --list --regex 'diffusion|transformer' --head 2 \
     --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### Parse and validate decision strings

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Converts a comma-separated string of decisions into an ordered, unique list. Handles the 'all' token.

```python
from openreview_downloader.cli import parse_decisions

print(parse_decisions("oral,all,accepted"))
# ['oral', 'accepted', 'rejected']

print(parse_decisions("spotlight,rejected"))
# ['spotlight', 'rejected']
```

--------------------------------

### Search with Multiple Text Criteria

Source: https://github.com/mireklzicar/openreview_downloader/blob/main/README.md

Require multiple terms or patterns by repeating the `--grep` (or `--search`) flag. This ensures that all specified terms are present in the matched papers.

```bash
ordl accepted --list --grep diffusion --grep protein --venue-id NeurIPS.cc/2025/Conference
```

--------------------------------

### compile_regexes(patterns, case_sensitive)

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Compiles a list of regex patterns into `re.Pattern` objects. This function is useful for preparing search patterns that can be efficiently applied to filter data.

```APIDOC
## compile_regexes(patterns, case_sensitive)

### Description
Compile regex patterns into `re.Pattern` objects. Raises `argparse.ArgumentTypeError` on invalid patterns.

### Parameters
- **patterns** (list[str]): A list of regex pattern strings to compile.
- **case_sensitive** (bool): If True, the regex matching will be case-sensitive. If False, it will be case-insensitive.

### Returns
A list of compiled `re.Pattern` objects.
```

--------------------------------

### Python API: decision_counts()

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Counts papers based on their decisions (oral, spotlight, accepted, rejected) from provided lists of accepted and rejected notes.

```APIDOC
### `decision_counts(accepted, rejected, venue_id)` — Count papers per decision

Returns a dict with keys `oral`, `spotlight`, `accepted`, `rejected`.

```python
from openreview_downloader.cli import (
    build_client, fetch_notes, decision_counts
)

client = build_client()
venue_id = "ICLR.cc/2025/Conference"
accepted, rejected = fetch_notes(client, venue_id, need_rejected=True)
counts = decision_counts(accepted, rejected, venue_id)
# {'oral': 103, 'spotlight': 322, 'accepted': 1734, 'rejected': 4812}
print(counts)
```
```

--------------------------------

### filter_selected(selected, args)

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Applies search and regex filters to a list of selected papers. It takes the output from `collect_selected` and a parsed `args` namespace to return only the papers that match the specified criteria, along with per-paper match metadata.

```APIDOC
## filter_selected(selected, args)

### Description
Takes the output of `collect_selected` plus a parsed `args` namespace and returns `[(note, category, path, match_info), ...]` with only matching papers, including per-paper match metadata.

### Parameters
- **selected**: The output from `collect_selected`.
- **args**: A parsed `argparse.Namespace` object containing filtering criteria such as `search_terms` and `regexes`.

### Returns
A list of tuples, where each tuple contains `(note, category, path, match_info)` for the matching papers.
```

--------------------------------

### parse_decisions(raw)

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

Parses a raw string of comma-separated decisions into an ordered, deduplicated list. It also handles the special token 'all' by expanding it to include both 'accepted' and 'rejected' decisions.

```APIDOC
## parse_decisions(raw)

### Description
Parse and validate decision strings. Converts a comma-separated decision string into an ordered, deduplicated list. Expands the `all` token into `["accepted", "rejected"]`.

### Parameters
- **raw** (str): A comma-separated string of decisions (e.g., "oral,all,accepted").

### Returns
An ordered, deduplicated list of decision strings.
```

--------------------------------

### Count papers per decision

Source: https://context7.com/mireklzicar/openreview_downloader/llms.txt

The `decision_counts` function calculates the number of papers for each decision category (oral, spotlight, accepted, rejected) given lists of accepted and rejected notes.

```python
from openreview_downloader.cli import (
    build_client, fetch_notes, decision_counts
)

client = build_client()
venue_id = "ICLR.cc/2025/Conference"
accepted, rejected = fetch_notes(client, venue_id, need_rejected=True)
counts = decision_counts(accepted, rejected, venue_id)
# {'oral': 103, 'spotlight': 322, 'accepted': 1734, 'rejected': 4812}
print(counts)
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.