Resolve conflicts

.docker/app_dockerfile

@@ -31,6 +31,7 @@ ARG VUE_APP_LOGO_URL=magic-logo-url
 ARG VUE_APP_HOMEPAGE_URL=magic-homepage-url
 ARG VUE_APP_EDITABLE_INVENTORY=magic-setting
 ARG VUE_APP_WEBSITE_TITLE=magic-title
+ARG VUE_APP_QR_CODE_RESOLVER_URL=magic-qr-code-resolver-url
 
 COPY webapp ./
 RUN /node_modules/.bin/vue-cli-service build

.docker/app_entrypoint.sh

@@ -26,6 +26,7 @@ echo "  LOGO_URL: ${VUE_APP_LOGO_URL}"
 echo "  HOMEPAGE_URL: ${VUE_APP_HOMPAGE_URL}"
 echo "  EDITABLE_INVENTORY: ${VUE_APP_EDITABLE_INVENTORY}"
 echo "  WEBSITE_TITLE: ${VUE_APP_WEBSITE_TITLE}"
+echo "  QR_CODE_RESOLVER_URL: ${VUE_APP_QR_CODE_RESOLVER_URL}"
 echo ""
 echo "Patching..."
 
@@ -36,6 +37,7 @@ for file in $ROOT_DIR/js/app.*.js* $ROOT_DIR/*html; do
     sed -i "s|magic-homepage-url|${VUE_APP_HOMEPAGE_URL}|g" $file
     sed -i "s|magic-setting|${VUE_APP_EDITABLE_INVENTORY}|g" $file
     sed -i "s|magic-title|${VUE_APP_WEBSITE_TITLE}|g" $file
+    sed -i "s|magic-qr-code-resolver-url|${VUE_APP_QR_CODE_RESOLVER_URL}|g" $file
     done
 
 echo "Done!"

.gitignore

@@ -12,6 +12,8 @@ webapp/cypress/videos
 #starting material files
 greyGroup_chemInventory*
 
+pydatalab/Pipfile
+
 
 *.sublime-workspace
 *.sublime-project

INSTALL.md

@@ -91,7 +91,7 @@ To make use of this file:
     - Detailed instructions for installing `pipenv`, `pip` and Python itself can be found on the [`pipenv` website](https://pipenv.pypa.io/en/latest/install/#installing-pipenv).
     - We recommend you install `pipenv` from PyPI (with `pip install pipenv` or `pip install --user pipenv`) for the Python distribution of your choice (in a virtual environment or otherwise). `pipenv` will be used to create its own virtual environment for installation of the `pydatalab` package.
 1. Install the `pydatalab` package.
-    - Navigate to the `pydatalab` folder and run `pipenv install --dev`.
+    - Navigate to the `pydatalab` folder and run `pipenv sync --dev`.
         - The default Python executable on your machine must be 3.10+, otherwise this must be specified explicitly at this point).
         - This will create a `pipenv` environment for `pydatalab` and all of its dependencies that is registered within *this folder* only.
         - You can remove this environment to start fresh at any time by running `pipenv --rm` from within this directory.

docker-compose.yml

@@ -14,6 +14,8 @@ services:
       - VUE_APP_LOGO_URL
       - VUE_APP_HOMEPAGE_URL
       - VUE_APP_EDITABLE_INVENTORY
+      - VUE_APP_WEBSITE_TITLE
+      - VUE_APP_QR_CODE_RESOLVER_URL
     ports:
       - "8081:8081"
 
@@ -94,8 +96,6 @@ services:
     restart: unless-stopped
     networks:
       - backend
-    ports:
-      - "27017:27017"
 
 networks:
   backend:

pydatalab/pydatalab/apps/chat/blocks.py

@@ -79,11 +79,11 @@ class ChatBlock(DataBlock):
                 "input_cost_usd_per_MTok": 5.00,
                 "output_cost_usd_per_MTok": 15.00,
             },
-            "gpt-3.5-turbo": {
-                "name": "gpt-3.5-turbo",
-                "context_window": 16385,
-                "input_cost_usd_per_MTok": 1.50,
-                "output_cost_usd_per_MTok": 2.00,
+            "gpt-4o-mini": {
+                "name": "gpt-4o-mini",
+                "context_window": 128_000,
+                "input_cost_usd_per_MTok": 0.15,
+                "output_cost_usd_per_MTok": 0.60,
             },
             "gpt-4": {
                 "name": "gpt-4",
@@ -155,10 +155,11 @@ def render(self):
                 return
 
         if self.data.get("model") not in self.data.get("available_models", {}):
+            bad_model = self.data.get("model")
             self.data["error_message"] = (
-                f"Chatblock received an unknown model: {self.data.get('model')}"
+                f"Chatblock received an unknown or deprecated model: {bad_model}. Reverting to default model {self.defaults['model']}."
             )
-            return
+            self.data["model"] = self.defaults["model"]
 
         try:
             model_name = self.data["model"]
@@ -237,10 +238,10 @@ def _prepare_item_json_for_chat(self, item_id: str):
         item_filenames = {
             str(file["immutable_id"]): file["name"] for file in item_info.get("files", [])
         }
-        for block in item_info.get("blocks_obj", {}).values():
-            block.pop("bokeh_plot_data", None)
 
-            block_fields_to_remove = ["item_id", "block_id"]
+        big_data_keys = ["bokeh_plot_data", "b64_encoded_image"]
+        for block in item_info.get("blocks_obj", {}).values():
+            block_fields_to_remove = ["item_id", "block_id", "collection_id"] + big_data_keys
             [block.pop(field, None) for field in block_fields_to_remove]
 
             # nmr block fields to remove (need a more general way to do this)

pydatalab/pydatalab/routes/v0_1/items.py

@@ -3,7 +3,7 @@
 from typing import Dict, List, Optional, Set, Union
 
 from bson import ObjectId
-from flask import Blueprint, jsonify, request
+from flask import Blueprint, jsonify, redirect, request
 from flask_login import current_user
 from pydantic import ValidationError
 from pymongo.command_cursor import CommandCursor
@@ -644,6 +644,9 @@ def get_item_data(
            call its render function).
 
     """
+    redirect_to_ui = bool(request.args.get("redirect-to-ui", default=False, type=json.loads))
+    if refcode and redirect_to_ui and CONFIG.APP_URL:
+        return redirect(f"{CONFIG.APP_URL}/items/{refcode}", code=307)
 
     if item_id:
         match = {"item_id": item_id}
@@ -767,6 +770,9 @@ def get_item_data(
     # Must be exported to JSON first to apply the custom pydantic JSON encoders
     return_dict = json.loads(doc.json(exclude_unset=True))
 
+    if item_id is None:
+        item_id = return_dict["item_id"]
+
     # create the files_data dictionary keyed by file ObjectId
     files_data: Dict[ObjectId, Dict] = {
         f["immutable_id"]: f for f in return_dict.get("files") or []

pydatalab/pyproject.toml

@@ -15,7 +15,7 @@ classifiers = [
     "Topic :: Scientific/Engineering",
 ]
 
-requires-python = ">=3.10"
+requires-python = ">= 3.10, < 3.12"
 
 dependencies = [
     # Slightly confusing, but git dependencies cannot be included as extra by uv at the moment
@@ -60,7 +60,7 @@ server = [
     "bokeh ~= 2.4, < 3.0",
     "Flask ~= 3.0",
     "Flask-Login ~= 0.6",
-    "Flask-Cors ~= 4.0",
+    "Flask-Cors ~= 5.0",
     "Flask-Dance ~= 7.1",
     "Flask-PyMongo ~= 2.3",
     "Flask-Mail ~= 0.10",

pydatalab/requirements/requirements-all-dev.txt

@@ -62,7 +62,7 @@ contourpy==1.2.1
     # via matplotlib
 coverage==7.6.1
     # via pytest-cov
-cryptography==43.0.0
+cryptography==43.0.1
     # via paramiko
 cycler==0.12.1
     # via matplotlib
@@ -106,7 +106,7 @@ flask==3.0.3
     #   flask-pymongo
 flask-compress==1.15
     # via datalab-server (pyproject.toml)
-flask-cors==4.0.1
+flask-cors==5.0.0
     # via datalab-server (pyproject.toml)
 flask-dance==7.1.0
     # via datalab-server (pyproject.toml)

pydatalab/requirements/requirements-all.txt

@@ -50,7 +50,7 @@ cloudpickle==3.0.0
     # via dask
 contourpy==1.2.1
     # via matplotlib
-cryptography==43.0.0
+cryptography==43.0.1
     # via paramiko
 cycler==0.12.1
     # via matplotlib
@@ -89,7 +89,7 @@ flask==3.0.3
     #   flask-pymongo
 flask-compress==1.15
     # via datalab-server (pyproject.toml)
-flask-cors==4.0.1
+flask-cors==5.0.0
     # via datalab-server (pyproject.toml)
 flask-dance==7.1.0
     # via datalab-server (pyproject.toml)

pydatalab/scripts/pipenv-resync.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+# This script is a hack to keep a pipenv lock synced with our pip-compile files
+# It injects dynamic optional dependencies into our pyproject.toml (currently incompatible
+# with any non-dynamic dependencies in setuptools 74) and generates the lock file
+#
+
+TOP_DIR=$(git rev-parse --show-toplevel)
+# Strip comments
+sed -i '/^\s&*#/d' $TOP_DIR/pydatalab/requirements/*.txt
+cd $TOP_DIR/pydatalab && pipenv install -r requirements/requirements-all.txt && pipenv install -r requirements/requirements-all-dev.txt --dev
+# Reset Pipfile, which will now have all deps inlined
+ cd $TOP_DIR/pydatalab && git checkout requirements

pydatalab/tests/server/test_graph.py

@@ -22,6 +22,21 @@ def test_simple_graph(admin_client):
 
     child_2 = Sample(
         item_id="child_2",
+        refcode="test:TEST02",
+        synthesis_constituents=[
+            Constituent(item={"type": "samples", "item_id": "parent"}, quantity=None),
+        ],
+    )
+
+    child_3 = Sample(
+        item_id="child_3",
+        synthesis_constituents=[
+            Constituent(item={"type": "samples", "item_id": "parent"}, quantity=None),
+        ],
+    )
+
+    child_4 = Sample(
+        item_id="child_4",
         synthesis_constituents=[
             Constituent(item={"type": "samples", "item_id": "parent"}, quantity=None),
         ],
@@ -45,25 +60,30 @@ def test_simple_graph(admin_client):
         cell_format="swagelok",
     )
 
-    new_samples = [json.loads(d.json()) for d in [parent, child_1, child_2, missing_child, cell]]
+    new_samples = [
+        json.loads(d.json())
+        for d in [parent, child_1, child_2, child_3, child_4, missing_child, cell]
+    ]
 
-    response = admin_client.post(
+    sample_list = admin_client.post(
         "/new-samples/",
         json={"new_sample_datas": new_samples},
     )
 
-    assert response.status_code == 207
-    assert all(d == 201 for d in response.json["http_codes"])
+    assert sample_list.status_code == 207
+    assert all(d == 201 for d in sample_list.json["http_codes"])
 
     graph = admin_client.get("/item-graph").json
     assert {n["data"]["id"] for n in graph["nodes"]} == {
         "parent",
         "child_1",
         "child_2",
+        "child_3",
+        "child_4",
         "missing_child",
         "abcd-1-2-3",
     }
-    assert len(graph["edges"]) == 4
+    assert len(graph["edges"]) == 6
 
     response = admin_client.post("/delete-sample/", json={"item_id": "missing_child"})
     assert response
@@ -73,18 +93,20 @@ def test_simple_graph(admin_client):
         "parent",
         "child_1",
         "child_2",
+        "child_3",
+        "child_4",
         "abcd-1-2-3",
     }
 
-    assert len(graph["edges"]) == 3
+    assert len(graph["edges"]) == 5
 
     graph = admin_client.get("/item-graph/child_1").json
     assert len(graph["nodes"]) == 2
     assert len(graph["edges"]) == 1
 
     graph = admin_client.get("/item-graph/parent").json
-    assert len(graph["nodes"]) == 4
-    assert len(graph["edges"]) == 3
+    assert len(graph["nodes"]) == 6
+    assert len(graph["edges"]) == 5
 
     collection_id = "testcoll"
     collection_json = {
@@ -107,5 +129,32 @@ def test_simple_graph(admin_client):
     assert len(graph["edges"]) == 2
 
     graph = admin_client.get("/item-graph/parent").json
+    assert len(graph["nodes"]) == 7
+    assert len(graph["edges"]) == 10
+
+    samples = sample_list.json["responses"]
+
+    refcode_child_3 = None
+    refcode_child_4 = None
+
+    for item in samples:
+        if item["item_id"] == "child_3":
+            refcode_child_3 = item["sample_list_entry"].get("refcode")
+        elif item["item_id"] == "child_4":
+            refcode_child_4 = item["sample_list_entry"].get("refcode")
+
+    response = admin_client.post(
+        f"/collections/{collection_id}",
+        json={"data": {"refcodes": [refcode_child_3, refcode_child_4]}},
+    )
+
+    assert response.status_code == 200
+    assert response.json["status"] == "success"
+
+    graph = admin_client.get(f"/item-graph?collection_id={collection_id}").json
     assert len(graph["nodes"]) == 5
-    assert len(graph["edges"]) == 8
+    assert len(graph["edges"]) == 4
+
+    graph = admin_client.get("/item-graph/parent").json
+    assert len(graph["nodes"]) == 7
+    assert len(graph["edges"]) == 14

pydatalab/tests/server/test_items.py

@@ -2,13 +2,19 @@ def test_single_item_endpoints(client, inserted_default_items):
     for item in inserted_default_items:
         response = client.get(f"/items/{item.refcode}")
         assert response.status_code == 200, response.json
+        assert response.json["item_id"] == item.item_id
+        assert response.json["item_data"]["item_id"] == item.item_id
         assert response.json["status"] == "success"
 
         test_ref = item.refcode.split(":")[1]
         response = client.get(f"/items/{test_ref}")
         assert response.status_code == 200, response.json
+        assert response.json["item_id"] == item.item_id
+        assert response.json["item_data"]["item_id"] == item.item_id
         assert response.json["status"] == "success"
 
         response = client.get(f"/get-item-data/{item.item_id}")
         assert response.status_code == 200, response.json
         assert response.json["status"] == "success"
+        assert response.json["item_id"] == item.item_id
+        assert response.json["item_data"]["item_id"] == item.item_id