832401304 Chen Qizhen | YOULIAN Address Book: Pair Development and Cloud Delivery
| Item | Content |
|---|---|
| Course for This Assignment | 2601_FZU-MU_SE |
| Author of this post | Chen Qizhen (陈颀桢, 832401304) |
| Team Name | 832401301蔡奇真832401304陈颀桢 |
| Assignment Requirements | Extreme Programming |
| Objectives of This Assignment | Implement address book features and practice pair development, Git and software engineering processes |
| Other References | Vue · FastAPI · SQLite · openpyxl · Horizon UI |
YOULIAN Address Book: Pair Development, Implementation and Cloud Delivery
目录
- YOULIAN Address Book: Pair Development, Implementation and Cloud Delivery
- 1. Assignment overview and requirement mapping
- 2. Project links and reproduction
- 3. Implementation design
- 3.1 Module boundaries and runtime structure
- 3.2 Data model and basic operations
- 3.3 Favorite contacts
- 3.4 Multiple types and repeated types of contact methods
- 3.5 Excel export and import
- 3.6 Visual style and error handling
- 4. Git contribution evidence and commit counts
- 4.1 Fixed-revision counting scope
- 4.2 Original GitHub commit-history screenshot
- 4.3 Code and collaboration evidence
- 5. Screenshots and complete demonstration video
- 5.1 Complete running video
- 5.2 Desktop overview
- 5.3 Multiple-method form and edited result
- 5.4 Favorites and quick access
- 5.5 Excel template, successful import and errors
- 5.6 Mobile layout and deletion confirmation
- 5.7 Supplemental full label round trip
- 6. Responsibilities, contribution ratio and peer assessment
- 7. Collaboration difficulties and solutions
- 8. Testing, cloud deployment and known boundaries
- 8.1 Verification results
- 8.2 Public deployment and reboot evidence
- 8.3 Boundaries and future improvements
- 9. Separate PSP tables for both members
- 9.1 Cai Qizhen (蔡奇真, 832401301)
- 9.2 Chen Qizhen (陈颀桢, 832401304)
- 10. Project reflection
- 10.1 From separate features to a complete workflow
- 10.2 Cai Qizhen: reflections on frontend development
- 10.3 Chen Qizhen: reflections on backend and data processing
- 10.4 Lessons from pair collaboration and Git
- 10.5 Lessons from testing and public deployment
- 10.6 Limitations and future improvements
1. Assignment overview and requirement mapping
This project is the Extreme Programming pair assignment for EE308FZ Software Engineering. We built YOULIAN, an address book with an English interface implemented in Vue 3, backed by FastAPI, SQLite and openpyxl for APIs, persistence and Excel processing. Cai Qizhen (蔡奇真, 832401301) handles the frontend. Chen Qizhen (陈颀桢, 832401304) handles the backend and contributes to integration, deployment and delivery.
This shared project record includes a separate PSP table for each member. Each member must publish an individual post using their own CSDN account and submit the URL of that post on the course assignment page.
| Assignment requirement | Implementation and evidence |
|---|---|
| Markdown, basic information, contents and links | Opening information table, contents, Section 2 project links and references linked in the opening table |
| Design, screenshots and responsibilities (30 basic requirement points in total) | Sections 3, 5 and 6; repository standards and verification documents |
| Page style (10 points) | Lavender, mint and cream-gold palette, phonebook illustration, desktop and mobile layouts |
| Favorite contacts (25 points) | Stars, favorite count and filter; persisted favorite state |
| Multiple contact methods (10 points) | phone, email, social and address; multiple entries of the same type |
| Excel import/export (25 points) | Genuine .xlsx, one contact per row, six categorized columns, templates and row-level errors |
| Cloud deployment (optional, up to 15 extra points) | Public Tencent Cloud access, service restart and whole-server reboot verification |
| Git, Release and PR | Substantial code commits by both members, a fixed 15-commit snapshot, v1.0.1 and merged PRs #1 and #2 |
| Complete video, contribution ratio and PSP | Full MP4 in Section 5, 50% each in Section 6 and two independent tables in Section 9 |
The base score is 100, with cloud deployment assessed separately as a bonus.
2. Project links and reproduction
| Resource | Link |
|---|---|
| Team GitHub repository | Source code |
| Coding standards | code-style.md |
| Responsibilities and API contract | division.md · api.md |
| Public address book | YOULIAN |
| Public health check | api/health |
| Full project Release | v1.0.1 |
| Backend collaboration PR | PR #1, merged |
| Frontend and integration PR | PR #2, merged |
| Excel format and examples | samples/README.md |
| Testing, integration and cloud verification | Backend tests · Integration · Cloud report |
| CSDN video (review approved) | Watch online |
| Full video and screenshot materials | Release materials |
The runtime uses Python 3.12 and Node.js meeting Vite requirements: 20.19+ within the 20.x line, or 22.12 and later. Dependencies are installed with the repository's pnpm lockfile.
git clone https://github.com/Cai5133/832401301-CaiQizhen-832401304-ChenQizhen.git
cd 832401301-CaiQizhen-832401304-ChenQizhen
cd backend
python -m venv .venv
# Linux / macOS
.venv/bin/python -m pip install -r requirements.txt
.venv/bin/python -m uvicorn app.main:app --host 127.0.0.1 --port 8000
On Windows, use these virtual-environment commands in PowerShell instead:
.\.venv\Scripts\python.exe -m pip install -r requirements.txt
.\.venv\Scripts\python.exe -m uvicorn app.main:app --host 127.0.0.1 --port 8000
In another terminal, enter the frontend directory from the repository root:
cd frontend
pnpm install --frozen-lockfile
pnpm dev
The development page is http://127.0.0.1:5173/; Vite proxies /api to http://127.0.0.1:8000. Tables are created on first startup, and an empty database produces an empty list. For production, run pnpm build first and then start the backend so FastAPI serves both the Vue build and the API. These localhost URLs are for local reproduction; the public URL in the table is the cloud deployment evidence.
3. Implementation design
3.1 Module boundaries and runtime structure
Before implementation, we separated page, API and data-processing responsibilities. The frontend owns presentation and interaction; the backend owns final validation and persistence. The shared docs/api.md defines fields, status codes, upload parameters and the Excel format.
| Directory / module | Responsibility |
|---|---|
| frontend/src/App.vue | Page state, contact selection, search, favorites and operation coordination |
| ContactList / ContactDetail | Contact list, stars and multiple-method details |
| ContactForm / BaseDialog | Dynamic form, validation, dialogs and focus management |
| ImportDialog / DeleteDialog | Excel results, row errors and deletion confirmation |
| frontend/src/api/contacts.js | Shared requests, timeouts, JSON errors and downloads |
| backend/app/schemas.py | Request models, types and length limits |
| backend/app/services/contacts.py | Contact queries, transactional writes and favorites |
| backend/app/services/excel.py | Workbook generation, parsing, batch validation and import |
| backend/app/database.py | SQLite connections, schema and a stable database path |
| deploy / docs / samples | Deployment configuration, collaboration records, evidence and fictional examples |
Browser: Vue page
| HTTP /api
v
Nginx :80 -> FastAPI / Uvicorn 127.0.0.1:8000
|-- Vue production files
|-- contact / favorite APIs -> SQLite
`-- Excel APIs -> openpyxl -> SQLite transaction
3.2 Data model and basic operations
The contacts table stores id, name and is_favorite. The contact_methods table references a contact through contact_id and stores type, label and value. This one-to-many relationship does not impose uniqueness on method type, so a contact may have two phone numbers or several email addresses.
Creation, editing and deletion use real APIs. An edit updates the contact, removes the previous methods and inserts the replacement methods in one transaction, preventing accumulation after repeated edits. Failures roll back the transaction. Foreign-key cascading removes methods when a contact is deleted. Phone numbers remain strings to preserve leading zeros, international prefixes and extensions. SQL uses parameters, and search text is matched literally.
3.3 Favorite contacts
Stars in the list and detail panel clearly identify favorites. Favorites in the sidebar provides quick access, while a statistic card displays the count. The frontend sends PATCH /api/contacts/{id}/favorite with an explicit isFavorite: true or false, rather than asking the backend to blindly toggle the value. The backend persists the state in SQLite and returns the full contact; the frontend updates its display only after success.
Removing a favorite while viewing the favorites filter removes that contact from the result and adjusts the selected detail. An empty result shows an empty state. Refreshes and service restarts reload the persisted value. The backend also supports the favorite=true query; the current frontend filters its loaded list locally.
3.4 Multiple types and repeated types of contact methods
The form dynamically adds and removes methods of type phone, email, social or address. Each entry contains an optional label and a required value. One contact can therefore have two phone numbers, an email, a social account and an address, totaling five methods.
Frontend and backend agree on a 1–100-character name, a 1–2000-character method value, a label of at most 100 characters and a maximum of 100 methods per contact. Unfilled new rows are ignored on submission; a label without a value produces an error. The frontend performs basic email-format validation. The list searches names and any method value, and the backend provides equivalent name/method queries.
3.5 Excel export and import
openpyxl creates genuine .xlsx files with a Contacts worksheet and one contact per row. The interface is English, while the workbook retains the six agreed Chinese headers to preserve template compatibility.
| 姓名 | 收藏 | 电话 | 邮箱 | 社交账号 | 地址 |
|---|---|---|---|---|---|
| Demo Alex | 1 | ["01000123456","+44 20 0000 0000 ext.9"] | ["demo@example.com"] | ["demo_social"] | ["12 Example Road, Block A"] |
These headers mean Name, Favorite, Phone, Email, Social account and Address. Favorite is 0 or 1. Each method column contains a JSON array of strings or objects with label and value; a blank cell or [] means no methods of that type. Multiple values remain in one categorized column, and commas within addresses are not split incorrectly. Export explicitly includes all contacts, regardless of the current search or favorites filter. Text cells are explicitly strings so values beginning with an equals sign are not interpreted as formulas.
Import proceeds as follows: download the template → select or drop an .xlsx → check extension, empty file and 5 MB limit on the frontend → upload using the multipart file field → validate worksheet, headers, types and every row on the backend → write the entire batch in one transaction → return the imported count and refresh the list. Any row error returns its row, column and reason with no records written. Database write failures also roll back the batch.
Additional backend limits are 50 MB of expanded ZIP data, at most 1000 ZIP entries, 5000 contacts per import and 10000 scanned data rows. Formulas, Excel error values and nonempty extra columns are rejected. Unreadable workbooks return 400, limit violations 413 and format or row errors 422. The frontend checks the ZIP signature on download to avoid saving an error page as an Excel file.
Complete business information: Excel v1.1 keeps the same six headers and accepts legacy string arrays. A labeled method is exported as {"label":"Work","value":"00123"}; unlabeled methods may remain strings. Import restores names, favorites, method types, labels and values. Duplicate names are permitted, and importing a file again adds records rather than deduplicating or updating existing contacts. Export rejects aggregated cells exceeding Excel's 32767-character limit instead of silently truncating them. These documented boundaries are possible future improvements.
3.6 Visual style and error handling
The page uses a light background, card hierarchy and purple accent, with lavender, mint and cream-gold for the address-book context. The phonebook character is a project SVG. References inform visual organization; business features follow the assignment. Sources and licenses are retained in the repository.
The interface is English and provides selected, loading, empty and error states. Failed saves preserve inputs, and busy operations prevent duplicate submissions. Deletion requires confirmation with initial focus on keeping the contact. Closing a modified form prompts about unsaved changes. Timeouts do not produce false success, and failed refreshes retain the previous list with a warning. Dialogs support keyboard use and focus management; narrow screens use a vertical layout.
4. Git contribution evidence and commit counts
4.1 Fixed-revision counting scope
The count was checked on November 16, 2026 and fixed at main commit 6429224280016a3d734c9fa841417c74553d496d. View its complete history. Commits are grouped by their linked GitHub author identity and include initialization, documentation and merges. A commit shared by several branches is counted only once.
| Member | GitHub identity | Including merges | Excluding merges |
|---|---|---|---|
| Cai Qizhen (蔡奇真, 832401301) | Cai5133 | 4 | 4 |
| Chen Qizhen (陈颀桢, 832401304) | grrrced; the Git author name also appears as 陈颀桢 | 11 | 8 |
| Total | Two actual contributors | 15 | 12 |
Both members have substantial code contributions, and the repository exceeds the three-commit minimum. Commit count is not a workload percentage.
4.2 Original GitHub commit-history screenshot
The image below directly captures GitHub's Commits page at 6429224280016a3d734c9fa841417c74553d496d, matching the 15 commits in the table. It is hosted on CSDN. Earlier summary images and API records remain in the original materials for cross-checking.

The Excel label fix in commit 6429224 is included in this snapshot. Three merge commits are included; commit counts are not converted directly into workload percentages.
4.3 Code and collaboration evidence
| Commit / PR | Content |
|---|---|
| 0242ba0 · Cai Qizhen | Frontend structure and initial page scaffold |
| 0d0d503 · Cai Qizhen | Complete frontend workflows, bright English interface, tests and acceptance documents |
| 5e684b8 · Chen Qizhen | Transactional contact/favorite APIs and persistence |
| 7c9470a · Chen Qizhen | Validated Excel import/export and examples |
| 77106b6 · Chen Qizhen | Production integration and cloud deployment preparation |
| PR #1 / PR #2 | Backend and frontend/integration deliveries, both merged |
The counts can be reproduced in the repository:
git fetch origin
git log --oneline 6429224280016a3d734c9fa841417c74553d496d
git rev-list --count 6429224280016a3d734c9fa841417c74553d496d
git rev-list --count --no-merges 6429224280016a3d734c9fa841417c74553d496d
git shortlog -sne 6429224280016a3d734c9fa841417c74553d496d
Cai Qizhen created the team repository. Chen Qizhen forked it under grrrced and opened PR #1, which was merged back into team main. The Fork, both merged PRs and the final Release are publicly verifiable. The course post provides no other designated upstream URL.
5. Screenshots and complete demonstration video
5.1 Complete running video
Watch the complete video on CSDN
Download the full MP4 (about 1 minute 55 seconds, Chinese subtitles) · Updated video, images and bilingual manuscripts
The complete recording has passed CSDN's review and is embedded below as video 553607. The CSDN viewing link and MP4 download above provide alternative access.
The recording accesses the actual Tencent Cloud Vue page and FastAPI / SQLite with fictional contacts, Chinese chapter captions and no narration. The first approximately 91 seconds retain the v1.0.0 core workflows, mobile layout and persistence demonstration. The final approximately 24 seconds add the v1.0.1 full label round trip. Both stages were actually recorded.
Chapter times come from the supplied subtitles:
| Start | Operation |
|---|---|
| 00:00 | Cloud project and page |
| 00:02 | Create a contact with five methods |
| 00:18 | Search by a method value |
| 00:21 | Edit the name and second phone number |
| 00:28 | Favorite, filter and remove a favorite |
| 00:35 | Export Excel |
| 00:37 | Download a template and select an import file |
| 00:43 | Row error and whole-batch rejection |
| 00:48 | Import four sample contacts |
| 00:56 | Mobile page and creation form |
| 01:03 | Restart the backend and check persistence |
| 01:10 | Confirm deletion |
| 01:15 | Restore from the exported Excel file |
| 01:21 | Import the valid sample again |
| 01:26 | Original demonstration result |
| 01:31 | Supplement: create three labeled methods |
| 01:40 | Export, delete the test contact and restore one row |
| 01:49 | Verify Personal, Work and School labels and values |
5.2 Desktop overview
The overview, statistics, list and detail panel share a light card-based style.

5.3 Multiple-method form and edited result
The form stores two phones, an email, a social account and an address; all five remain after editing.

The edited contact details show the name, favorite state and all methods.

5.4 Favorites and quick access
Stars, favorite count and the Favorites filter remain consistent; removing a favorite updates the result.

5.5 Excel template, successful import and errors
The import dialog provides a template, file selection and format instructions.

A valid sample reports four imported contacts.

An invalid workbook reports malformed JSON in the Phone column on row 2, with no batch writes.

5.6 Mobile layout and deletion confirmation
At 390px, content stacks vertically while keeping the main operations usable.

The mobile form keeps fields, buttons and scrolling accessible.

Deletion requires confirmation before removing the record.

5.7 Supplemental full label round trip
The public webpage creates three methods labeled Personal, Work and School. The actual exported workbook is reduced to the dedicated test contact's single row. After deleting that contact through the page, the workbook restores it through the import dialog. API checks confirm identical method types, labels and values, including leading zeros. Only dedicated test contacts are removed; the five preexisting contacts retain identical complete JSON.

These screenshots capture different stages. Counts change with creation, import, filtering and deletion; the images do not represent one simultaneous database state.
6. Responsibilities, contribution ratio and peer assessment
| Member | Main implementation and deliverables | Contribution |
|---|---|---|
| Cai Qizhen (蔡奇真, 832401301) | Vue components and bright English UI, dynamic multi-method form, detail/list, favorite/search interactions, Excel upload/download interface, request wrapper, frontend errors, mobile adaptation, frontend tests and handoff documents | 50% |
| Chen Qizhen (陈颀桢, 832401304) | FastAPI, SQLite schema and transactions, contact/favorite APIs, Excel generation/parsing and batch validation, backend tests, production integration, deployment configuration, cloud/restart verification, demonstration materials and backend documents | 50% |
| Total | Shared planning, API agreement, collaboration delivery, issue review and blog preparation | 100% |
The 50%/50% split supplied for this draft assesses functional scope and implementation work, rather than converting commit counts into percentages. Frontend work is concentrated in fewer commits; backend history includes separate feature, integration, merge and evidence-document stages. Different counts therefore do not require different contribution ratios.
Peer assessment: Cai Qizhen delivered the operation entry points, visual revisions and complex form feedback that make the required features usable. Chen Qizhen delivered data reliability, Excel processing, integration and public operation so results persist. The two parts depend on each other and each has verifiable code commits. Both can improve by freezing API boundaries earlier and recording PSP time at each stage, reducing retrospective reconstruction.
7. Collaboration difficulties and solutions
The following issues are grounded in the interface revisions, API contract and integration report. Preventive boundary handling is not presented as an undocumented production incident.
| Main owner | Issue / difficulty | Solution and result |
|---|---|---|
| Cai Qizhen | The initial dark interface did not match the requested bright style and required layout/color revisions | Used open-source dashboard and supplied image references; adopted lavender, mint and cream-gold, added a phonebook SVG, unified English text and retained source/license records |
| Cai Qizhen | Forms, favorites and Excel interaction needed progress before the real backend was connected | Checked interactions against an isolated fictional service following the contract, then integrated the real backend; preview results were not treated as database persistence evidence |
| Both | Early frontend limits differed from backend integration limits, risking rejection when editing otherwise valid imported content | Unified value, label and method limits to 2000 characters, 100 characters and 100 entries, and updated the contract/tests; the final frontend has 16 passing tests |
| Cai Qizhen | Failures, timeouts and refresh errors require preserved inputs and understandable state | Kept failed forms open, retained old data on refresh errors and blocked duplicate busy submissions; a simulated 503 checked input preservation |
| Chen Qizhen | Repeated method types, leading zeros and commas in addresses need unambiguous storage | Used a one-to-many method table and string values. Excel arrays preserve all values; labeled methods use label/value objects in v1.1. Round-trip checks compare the restored contents |
| Chen Qizhen | Partial import failures could leave half a batch, while changing working directories could select another database | Validated the full batch before a transactional write with rollback; fixed the database path and checked contacts/favorites after a real process restart |
| Both; integration led by Chen Qizhen | Shared README and handoff documents conflicted during merging | Reconciled shared content against both implementations, preserved code/history and verified the production build against the real backend |
| Chen Qizhen | Local operation alone cannot establish public availability or restart persistence | Deployed Nginx, systemd and a separate data directory; verified service restart, whole-server reboot, automatic recovery and identical complete data for five contacts |
| Both | Earlier materials and the later fix have different commit counts | Captured the actual GitHub page again and froze the latest count at 6429224: 15 commits; earlier materials retain their original snapshot labels |
For collaboration, API or Excel changes first update docs/api.md and examples, followed by both implementations. Shared-file conflicts are resolved by content rather than overwriting a teammate's files. PRs preserve reviewable delivery scopes. Final acceptance uses real APIs and a production build, not simply a page that opens.
8. Testing, cloud deployment and known boundaries
8.1 Verification results
This section summarizes repository reports without mixing stages. The initial backend delivery passed 40 tests and the first integrated version passed 42. The label fix and four additional invalid-object cases bring the current backend count to 46. The final integrated frontend passed 16 tests and its production build. See the integration report for scope.
| Verification | Recorded result |
|---|---|
| Frontend validation and utility tests | 16 passing tests; production build passed |
| Backend API, persistence, Excel and static routes | 46 passing tests |
| Clean-environment reproduction | Initial backend: 40 tests and real HTTP checks passed in a new directory/environment; frontend built after clean installation from its lockfile |
| Complete browser workflows | Five-method creation, editing, social-account search, favorite/removal and deletion passed |
| Excel round trip and failures | Actual downloads, template, valid import, row rejection, empty-database restoration and leading zeros passed; the supplemental check also preserves labels |
| Responsiveness and errors | No horizontal overflow at 390px in page or form; acceptance records report no page script errors |
| Persistence | Data survived a real backend process restart; complete JSON for five contacts matched after a cloud-server reboot |
# From the repository root (Linux / macOS)
cd frontend
pnpm test
pnpm build
cd ../backend
.venv/bin/python -m pip install -r requirements-dev.lock.txt
.venv/bin/python -m pytest -q
.venv/bin/python scripts/smoke_http.py
8.2 Public deployment and reboot evidence
The public URL is http://43.139.206.206/. It runs on a Tencent Cloud CVM in Guangzhou Zone 6 with Ubuntu 24.04, 2 CPU cores, 2 GB memory, a 50 GB system disk and 3 Mbps bandwidth. Nginx listens on public port 80 and proxies to local Uvicorn. systemd manages the youlian service. The database is /var/lib/youlian/contacts.sqlite3, separate from application release directories.
The November 16, 2026 cloud verification records cover complete browser workflows, Excel round trips, service restart and server reboot. During blog preparation, read-only requests to the homepage and /api/health again returned HTTP 200, with {"status":"ok"} from the health endpoint. This availability check does not replace the earlier functional acceptance.


The trial expires on December 16, 2026, 15:19 Beijing Time. Renewal is manual and automatic renewal is disabled. The URL is not a permanent hosting commitment and should be checked promptly for assessment. Public demonstrations use fictional contacts only.
8.3 Boundaries and future improvements
Excel v1.1 now preserves custom labels. Repeated imports still add records without deduplication, and the application has no pagination. Import previews, duplicate detection and pagination remain future improvements.
9. Separate PSP tables for both members
The supplied figures total 360 estimated and 360 actual minutes per member. They are retrospective estimates, not continuously timed phase records or evidence that an exact plan was registered before coding.
9.1 Cai Qizhen (蔡奇真, 832401301)
| PSP activity | Development phase | Estimated time (retrospective, min) | Actual time (retrospective estimate, min) |
|---|---|---|---|
| Planning | Planning and task allocation | 20 | 20 |
| Analysis | Requirement analysis and acceptance criteria | 25 | 25 |
| Learning | Required technical study and references | 10 | 10 |
| Coding Standard | Coding standards and API agreement | 15 | 10 |
| Design | Component / data model and detailed design | 30 | 35 |
| Coding | Implementation and deployment configuration | 130 | 135 |
| Code Review | Code review and corrections | 20 | 20 |
| Test | Testing, integration and acceptance | 50 | 45 |
| Reporting | Documents, demonstration and blog preparation | 30 | 35 |
| Measurement | Record review and workload accounting | 15 | 15 |
| Postmortem | Retrospective and improvement plan | 15 | 10 |
| Total | Total | 360 | 360 |
The frontend allocation emphasizes coding, dynamic forms and interaction. Detailed design, visual revisions and documentation are allocated slightly more time than planned, with a corresponding reduction in testing. These are retrospective allocation differences, not a basis for precise measured-efficiency claims.
9.2 Chen Qizhen (陈颀桢, 832401304)
| PSP activity | Development phase | Estimated time (retrospective, min) | Actual time (retrospective estimate, min) |
|---|---|---|---|
| Planning | Planning and task allocation | 20 | 20 |
| Analysis | Requirement analysis and acceptance criteria | 25 | 25 |
| Learning | Required technical study and references | 10 | 10 |
| Coding Standard | Coding standards and API agreement | 15 | 10 |
| Design | Component / data model and detailed design | 35 | 35 |
| Coding | Implementation and deployment configuration | 120 | 125 |
| Code Review | Code review and corrections | 20 | 20 |
| Test | Testing, integration and acceptance | 50 | 50 |
| Reporting | Documents, demonstration and blog preparation | 35 | 40 |
| Measurement | Record review and workload accounting | 15 | 15 |
| Postmortem | Retrospective and improvement plan | 15 | 10 |
| Total | Total | 360 | 360 |
The backend allocation covers APIs, database transactions, Excel processing and deployment configuration. Testing includes real HTTP, integration, service restart and server reboot checks. Documents and demonstration preparation receive slightly more time than planned. Each table totals 360 minutes independently; parallel work by two members must not be treated as the project's elapsed calendar duration.
10. Project reflection
10.1 From separate features to a complete workflow
This pair assignment took us through requirement analysis, task allocation, frontend and backend development, API integration, testing and cloud delivery. Although an address book has a relatively focused feature set, each operation affects other modules. Favorites, for example, require more than a star button: list markers, statistics, filters and stored state must agree. Multiple contact methods also require more than extra input fields. They involve dynamic addition and removal, validation, replacement during editing and preservation through Excel import and export.
After completing the project, we have a more concrete understanding of what it means to finish a feature. Users should be able to complete an operation, understand the reason when it fails and retrieve their data after a refresh or restart. Connecting these steps is what makes the address book a complete workflow.
10.2 Cai Qizhen: reflections on frontend development
My strongest lesson from frontend development is that visual presentation and interaction logic need to be considered together. Interface revisions involve more than colors and card styles. They also determine whether users can quickly find a contact, identify the current selection and understand the results of favoriting, importing and deleting. Through layout and color revisions, the project adopted a light background, layered cards and a phonebook character, while accommodating both desktop and mobile screens to make the main operations clearer and more focused.
The dynamic contact-method form was another important part. Creation, editing, blank rows, repeated types and invalid inputs each require appropriate handling. Checking successful submissions alone can miss details such as whether inputs survive a failed save or whether closing a form discards changes. I therefore began to consider loading, success, failure, empty results and unsaved changes as parts of one interaction, rather than simply implementing an API call for each button. This taught me that the frontend is responsible for the operation as it unfolds and for explaining the current state clearly to users.
10.3 Chen Qizhen: reflections on backend and data processing
Backend development strengthened my attention to data structures and reliable writes. The one-to-many relationship between contacts and methods supports different method types as well as several entries of the same type. Replacing the method list during editing and cascading related records during deletion help prevent leftover data and accumulation after repeated edits.
Excel processing also showed me that downloading or reading a file does not by itself establish correct import and export. Leading zeros, international phone numbers, multiple values of one type and commas in addresses require an explicit format to be preserved. Whole-batch validation and transactional writes ensure that an invalid file does not leave a partially imported batch. These choices helped us consider data formats, error feedback and persistence together. We also documented boundaries, including the original label omission, now fixed, and repeated imports adding records, to guide future improvements.
10.4 Lessons from pair collaboration and Git
Dividing frontend and backend responsibilities does not mean communication can wait until both parts are finished. API fields, length limits, response structures and Excel headers affect both implementations. Unifying field boundaries and resolving shared-document conflicts during integration taught us that API documentation must evolve with the code. Changes to shared agreements also need to reach the teammate promptly; otherwise, individually correct implementations can still become inconsistent when combined.
Git and PRs provided an inspectable record of collaboration. Commit history distinguishes each member's implementation, while preserved merge records show how the final version was assembled. The project also helped us understand that commit counts are only one part of the development record and do not directly represent workload. A more useful assessment considers functional scope, implementation difficulty, testing and delivery, and explains the responsibilities each member undertook.
10.5 Lessons from testing and public deployment
Testing turned an impression that the application worked into specific checkable results. Besides normal creation, editing and favorite workflows, we needed to examine invalid files, invalid inputs, failed requests and empty lists. Automated tests make rules easier to check repeatedly, while browser operations verify how the page and real APIs work together. Both serve a purpose.
Public deployment expanded the verification scope further. Starting a development service establishes local operation, but deployment also requires consistent access to static pages and APIs, a stable database location and recovery after service or server restarts. Real restart checks connected our persistence design to observed runtime results. They also helped us understand the role of setup instructions, deployment configuration and acceptance evidence in delivering a project.
10.6 Limitations and future improvements
There is still room for improvement. Excel v1.1 now preserves labels and values, while repeated imports still do not automatically deduplicate records. Future work can introduce import previews and duplicate warnings. Larger contact collections will also require attention to pagination and query efficiency to avoid loading too much data at once.
For the development process, we want to agree on API boundaries earlier, commit relatively independent tasks promptly and record time by stage from the start, giving planning and reflection a stronger basis. Future pair projects should continue to maintain coding standards, API descriptions, test records and deployment instructions so teammates can take over and reviewers can reproduce the work. Our main gain from this assignment is learning to connect interface implementation, data reliability and collaboration into a project that can be run, inspected and delivered.
相关推荐
企业数字化转型AI大模型赋能智能体平台及AI场景应用规划方案(生态合作版).pptx
企业数字化转型AI大模型赋能智能体平台及AI场景应用规划方案(生态合作版).pptx
MCGS的Http协议驱动,支持http协议
硬件配置 参数设置 设备属性 设备命令 通讯状态 附录 #HTTP 1、 设备简介 驱动类型 该驱动为独立设备,直接添加构件即可,不需要挂载到父设备下 通讯协议 采用标准HTTP1.1协议 通讯方式 驱动构件为客户端。 2、 硬件连接 本软件在通讯之前,请保证通讯连接正确。通讯连接方式:采用标准的RJ45接口的以太网接线(对调网线或使用HUB交换)。 3、 设备通讯参数 本设备为独立设备,无父设备参数 4、 设备构件参数设置 “HTTP设备”参数设置如下: <1> 内部属性: 构建url信息。 <2> 最小采集周期:本软件对设备进行操作的时间周期,单位为ms,默认为100ms。 <3> 通讯等待时间:通讯数据接收等待时间,默认设置为200ms,当采集速度要求较高或数据量较大时,设置值可适当减小或增大。 <4> HttpHead:http head添加需要的字段,不需要添加http head时值为空,默认不添加。 5、 设备命令 本设备构件提供设备命令,用于进行相应的操作,设备命令的格式如下: 设备命令 命令介绍 GetHttpHead 格式1 GetHttpHead(headStr) 用途 获取http head参数 参数 headStr 参数可以为字符型变量、字符串常量 例1 !SetDevice(设备0,6,"GetHttpHead(headStr)") 获取http head数据保存到headStr中 SetHttpHead 格式1 SetHttpHead(headStr) 用途 设置http head参数 参数 headStr 参数可以为字符型变量、字符串常量
印尼稻米收成与粮食安全数据集2026年
本数据集收录印度尼西亚2026年水稻收成与粮食安全遥测数据,包含10000条稻米农业收成产量与粮食安全监测记录,围绕印尼国家粮食安全体系中的核心作物水稻展开。稻米是印尼逾两亿人口的主粮,其收成波动直接影响国家粮食安全与进口政策,本数据集将种植区域、收成产量、气候与土壤相关遥测指标整合为结构化表格,可支撑粮食产量预测、收成影响因素分析、区域粮食安全评估与农业政策模拟等研究任务。典型应用场景包括:农业时间序列产量预测模型构建、气候因子与水稻收成的关联分析、省份与地区层面的粮食自给率测算、粮食安全风险预警指标设计,也可作为农业经济、粮食政策与数据科学交叉课程的教学素材。数据为2026年最新版,规模约1.2MB、单CSV结构规整,字段命名清晰,适合Python、Pandas、Scikit-learn等主流工具链直接使用,便于开展探索性分析、可视化与机器学习全流程练习。重要说明:本数据为遥测整理数据,分析结论以数据集文档口径为准,农业决策请结合官方统计综合研判。许可:CC0公有领域,可自由使用与二次分发。
智慧环保生态保护行业可信数据空间应用规划方案(技术方案版).pptx
智慧环保生态保护行业可信数据空间应用规划方案(技术方案版).pptx
气象服务行业可信数据空间应用规划方案(技术方案版).pptx
气象服务行业可信数据空间应用规划方案(技术方案版).pptx



